Professional Documents
Culture Documents
stream philosophy had run out . What had instead begun to seem
promising was the new wave in philosophical method , which ceased
to pander to " ordinary language" and which began in earnest to
reverse the antiscientific bias typical of " linguistic analysis." Even
here I had a major misgiving , however , because the sciences em -
did
not
include
neuroscience
. Indeed
, the best
of what
there
dualism - that dismissed neuroscience as largely irrelevant to theories in psychology and philosophy . Since I was a materialist and
hence
believed
that
the
mind
is the
brain
, it seemed
obvious
that
to know
how
we
see , how
we think
and
reason
and
decide
. I
Preface
the boundaries are ill defined . This book is thus the result of what I
came to regard as neurophilosophicalinquiries.
Given the range of topics I needed to know about, I was through out the project necessarily dependent on the willingness of neuroscientists to explain their research, to tell me what they thought was
important and why , and to give advice on who else to talk to and
what to read. My vvorst fear- that as a philosopher I would be considered an utter waste of time - was virtually never realized. Invari ably neuroscientists were exceedingly generous, often going beyond
explanations asked, allowing me to observe or participate in experiments, explaining details of techniques, and drawing back the curtain
on the wider vision that motivated their research.
From time to time I found considerable disagreement among
neuroscientists on fundamental issues, and at first I tacitly assumed
that there must be someone who really knew what was what and who
could settle for me what is the Truth . In the end I knew that I had to
make up my own mind , and do it the way any neuroscientist would :
find out as much as I reasonably could about the issue and go with
what seemed most reasonable. A vague decision procedure, to be
sure, but the only one I know of .
Without the generosity and patience of many people, not only
neuroscientists , but also philosophers , psychologists, and computer
scientists, this book would still be a shapeless intention . To Larry
Jordan I am especially indebted for giving me a basis in neurophysiol ogy and in laboratory techniques and for convincing me that it is
essential to think about how organisms move. I also owe an enormous debt to Rodolfo Llinas , whose unique blend of experimental
understanding and drive for theory gave me a sense of what a large,
unifying framework for neurobiology and psychology might look like
and how known data would figure in that framework . For similar
reasons, I am grateful to Francis Crick , whose general understanding
of the entire field of vision and sense of how theoretical problems in
neurobiology might be solved directed me in getting a grip on the
functional questions . More than anyone else, Llinas and Crick have
made neuroscience seem like the most exciting thing in the world and
consequently rekindled - and, I suspect, realigned- my philosoph ical preoccupations .
Among philosophers , my first and greatest debt is to Paul M .
Churchland , who has been a partner in the venture from the very
beginning . It was he especially who convinced me of the importance
of bringing science and the philosophy of science to bear on questions
in the philosophy of mind , and this has made all the difference in
thinking about consciousness, cognition , and subjective experience,
and about the general framework needed for a unified science of the
Preface
.
Xl
PSC
LaJolla1985
N europhilosophy
One ought to know that on the one hand pleasure, joy , laughter , and
games, and on the other, grief , sorrow , discontent, and dissatisfaction arise
only from the brain . It is especially by it that we think , comprehend, see,
and hear, that we distinfl,uish the ugly from the beautiful , the bad from the
good, the agreeablefrom the disagreeable. .
Hi ppocra tes
Philosophyis like the mother who gave birth to and endowedall the other
sciences
. Therefore
, one should not scorn her in her nakedness
and poverty,
but should hope, rather! that part of her Don Quixote ideal will live on in
her children so that they do not sink into philistinism.
Albert Einstein , 1932
GeneralIntroduction
General Introduction
empowered for all that to disemburden itself and to bootstrap its way
to insight and understanding .
It is within this context that certain intriguing problems ariseproblems concerning how to study the brain , how to conceive of
what it is up to, and how our commonsense conceptions of ourselves
might fit or fail to fit with what we discover. Some of these have
traditionally been recognized as philosophical problems . For example: Are mental states identical to brain states? Are mental states
reducible to brain states? What sort of business is reduction ? What are
emergent properties and are there any? What , if anything , is special
about the subjective point of view ? Are conscious experiences physio logically understandable ? What are representations and how can a
brain represent the world outside itself ?
Such philosophical questions are synoptic in character, in the sense
that they are very general and very broad . But they are not of an
entirely different nature from synoptic problems traditionally characterized as empirical : How is color vision produced ? How does the
brain learn and how does it store information ? What are representations and how does a brain represent the world outside itself ? Is the
hu~ an brain more complicated than it is smart?
The questions, whether asked by philosophers or by neuroscientists, are all part of the same general investigation , with some questions finding a natural home in both philosophy and neuroscience. In
any caseit is the same curiosity that bids them forth , and it is perhaps
best to see them all simply as questions about the brain and the
mind - or the mind -brain - rather than as questions for philosophy or
for neuroscience or for psychology . Administrative distinctions have a
purpose so far as providing office spaceand salaries is concerned, but
they should not dictate methods or constitute impedimenta to easy
exchange. This is not to deny that there are divisions of .laborindeed, within neuroscience itself there are divisions of labor- but it
is to argue that such divisions neither imply nor justify radical differ ences in methodology .
Philosophical problems were once thought to admit of a priori solutions, where such solutions were to be dredged somehow out of a
" pure reason," perhaps by a contemplation unfettered and uncontaminated by the grubbiness of empirical facts. Though a convenience
to those of the armchair persuasion, the dogma resulted in a rather
anti-intellectual and scoffing attitude toward science in general, and
when the philosophy was philosophy of mind , toward neuroscience
in particular . But with the publication in the 1960sof Quine 's Word
and Object and Sellars's Science
, Perceptionand Reality, it came to be
seen that philosophy at its best and properly conceived is continuous
General Introduction
with the empirical sciences, and that while problems and solutions
can be more or less synoptic , this is a difference in degree, not a
difference in kind . Although theories may be more or less distant
from observations, they are interesting only insofar as they can touch ,
finally , upon observations . Sometimes the route to observations may,
as in theoretical physics, be a long one through much theory , but a
route there must finally be.
What used to pass for a priori arguments about the impossibility of
science discovering this or that (such as the impossibility of discovering that space is non -Euclidean or that mental states are brain
states) were sometimes merely arguments based on what could or
could not be imagined by some individual philosopher . Since what
can or cannot be imagined about the empirical world is not indepen dent of what is already understood and believed about the empirical
world , failures of imaginability were all too often owed to ignorance
or to inflexible imaginations .
The sustaining conviction of this book is that top-do\-\Tn strategies
(as characteristic of philosophy , cognitive psychology , and artificial
intelligence research) and bottom -up strategies (as characteristic of
the neurosciences) for solving the mysteries of mind -brain function
should not be pursued in icy isolation from one another . What is
envisaged instead is a rich interanimation between the two , which
can be expected to provoke a fruitful co-evolution of theories, models,
and methods , where each informs , corrects, and inspires the other .
For neuroscientists , a sense of how to get a grip on the big questions and of the appropriate overarching framework with which to
pursue hands-on research is essential- essential, that is, if neuroscientists are not to lose themselves, sinking blissfully into the sweet,
teeming minutiae , or inching with manful dedication down a deadend warren . For philosophers , an understanding of what progress
has been made in neuroscience is essential to sustain and constrain
theories about such things as how representations relate to the world ,
whether representations are propositional in nature, how organisms
learn, whether mental states are emergent with respect to brain
states, whether conscious states are a single type of state, and so on.
It is essential, that is, if philosophers are not to remain boxed within
the narrow canyons of the commonsense conception of the world or
to content themselves with heroically plumping up the pillows of
decrepit dogma .
The guiding aim of the book is to paint in broad strokes the outlines
of a very general framework suited to the development of a unified
theory of the mind -brain . Additionally , it aims to bestir a yen for the
enrichment and excitement to be had by an interanimation of philo so-
General Introduction
General Introduction
philosophical research and insights in a coherent and readable fashion , trying to balance between providing sufficient detail to make
points thoroughly and being clear enough and clean enough so that
neuroscientists do not give up on it as painfully abstruse, or " philo sophical" in the bad sense of the word - that is, perverse, dark , and
anyhow pointless . Philosophical detail is apt to dissolve into mere
crinkum -crankum , and it is my intention to risk snubbing the niceties
in order to preserve an uncluttered pattern of the main arguments .
In the most straightforward sense, what is wanted is a unified
theory of how the mind -brain works . We want a theory of how the
mind -brain represents whatever it represents, and of the nature of
the computational processes underlying behavior . The collective effort to devise such a theory will be constrained by empirical facts at all
levels, including neurophysiological , ethological, and psychological
facts. In addition , it will be colored by pretheoretic hunches concerning what a theory could look like and what are the basic principles of
mind -brain operation . More fundamentally perhaps, it will also be
affected by opinions concerning whether such an enterprise is even
reasonable at all .
The idea that ultimately there should be a unified theory of the
brain - a theory that encompasses all levels of description - has of
course been around for a long time . But the idea has typically seemed
both surpassingly vague and pathetically remote. In truth , it really
has been less a palpable conception than a misty ideal toward which
science, in the very long haul , might progress. Consequently , philos ophy has tended to ignore developments in the neurosciences and
pretty much to go its own way . Likewise , research in the neurosciences has proceeded without much heed to what philosophers had to
say about the nature of knowledge or of mental states. Quite simply ,
neither found the other useful , and the two disciplines have had
largely independent histories . Contact was made only seldom, and
then it usually consisted in desultory sparring on the " mind -body
problem ."
But things are changing . Developments in neuroscience and in
philosophy , as well as developments in psychology and computer
science, have brought the disciplines to the stage where there are
common problems , and there is a gathering sense of the benefits for
research in cross-talk . For one thing , neuroscience has progressed to
the point where we can begin to theorize productively about basic
principles of whole brain function and hence to address the questions
concerning how the brain represents, learns, and produces behavior .
Second, many philosophers have moved away from the view that
philosophy is an a priori discipline in which philosophers can dis-
mechanisms might implement the functions . Fourth , work in computer science and computer modeling of networks has helped to
generate concepts of information processing, representation , and
computation that take us well beyond the earlier ideas and provide at
least a general sense of how to address the question of subintrospective
mind
- brain
processes
analysis, the better will be our brains' theoretical evolution . Thus, the
co-evolution of macrotheory and microtheory - broadly , of neuroscience and psychology - is a major methodological theme throughout .
The strategy shaping the book, therefore, has been to introduce phi losophy and neuroscience , each to the other .
Ideally , I should have liked to begin by presenting theoreticallyrelevant neuroscience . The problem in pursuing
neuroscience
. This
does
not
mean
that
there
are no
theories
General Introduction
by presenting some rudimentary neurophysiology , some rudimen tary neuroanatomy , a glimpse into neurology and into neuropsychol ogy, and a precis of a few methods used to study nervous systems.
I am painfully aware of how voluminously much I have left out of
my introduction to neuroscience, but my hope is twofold : that I have
presented enough so that philosophers may now approach textbooks
and review papers without being intimidated , and that I have said
enough so that the newly emerging theoretical frameworks presented
in chapter 10 can be understood . I present these frameworks as examples of what a large-scale theory of brain function might look like , but
at the same time I acknowledge that none has yet, and none might
ever, achieve the status of Governing Paradigm.
Philosophers who are expecting to find in the introduction to
neuroscience a point -by-point guide of just what facts in neuroscience
are relevant to just which traditional philosophical problems will be
disappointed . I have made some occasional efforts in that direction ,
but in the main my eye is on the overarching question of the nature of
a unified , integrating theory of how , at all its levels of description , the
brain works . If philosophers
to
account
for
what
we
think
we
know
about
the
nature
of
nature
of intertheoretic
reduction
General Introduction
10
General Introduction
very
abstract
considerations
. If I am
to defend
the
reasonableness
of
searching for a unified theory , I must answer these objections. Second , the theme of representations
most thoroughly in a philosophical context, though where the psychological sciences offer relevant principles and pertinent data, I try
to draw these in . Even so, the research in psychology and ethology is
insufficiently discussed, and this because a third and familiar practical reason began to assert itself : the book is already long enough .
It is difficult to resist the excitement that now typifies so much
research in the neurosciences and the related psychological sciences.
The excitement is generated in part because neuroscience is science
,
and in pushing back the bounds of darkness it is discovering surpris ing new things and teaching us how some aspect of the universe
works . But it is also because the discoveries have immediately
to do
anyone's philosophical aspirations , be they ancient or modern , un tutored or scholarly , as any quest there is.
Chapter 1
The Science of Nervous Systellls : A Historical
Sketch
1 .1
I n trod uction
If you root yourself to the ground , you can afford to be stupid . But if
you move , you must have mechanisms for moving , and mechanisms
to ensure that the movement is not utterly arbitrary and independent
of what is going on outside . Consider a simple protochordate , the sea
squirt . The newborn must swim about and feed itself until it finds a
suitable niche , at which time it backs in and attaches itself perma nently . Once attached , the sea squirt ' s mechanisms for movement
become excess baggage , and it wisely supplements its diet by feasting
on its smartest parts .
Animals are movers , and some of them display astonIshing agility .
How is it possible for an owl to dive , almost silently , out of the night
sky and to entrap a scurrying mouse in its talons ? Both organisms are
on the move , yet the owl ' s timing is precise , and it neither crashes
into the ground nor comes up empty -handed . How is it possible
simply to walk , and to walk at varying speeds and over sundry obsta cles? Look at a nervous system that is not performing normally because it has been altered by drugs , or by disease , or by trauma to the
inner ear , for example , and we get a glimpse of the awesome com plexity that underlies the smooth coordination we standardly take for
granted . What is going on inside a canary when it learns the motor
skill for song production , or inside wolves when they know how to
organize themselves to bring down a deer ? How is it that we see,
hear , and figure things out ?
Neurons are excitable cells , and neurons on the sensory periphery
are activated by such things as photons or vibration , while neurons
on the motor periphery cause the contraction of muscles . In between
14
SomeElementary
Neuroscience
are neurons
that
orchestrate
the
sequence
of muscle
cell
contractions
from
earlier
kinds
of brains
that
our
kind
of brain
was
not
built from scratch especially for us, but has capacities and limitations
that are due to its historical origins . The pressure for nervous systems
to evolve has derived not from the intrinsic beauty of rationality or
Brains are exceedingly complex and delicate, and they give up their
secrets with
exasperating
reluctance . Understanding
something
of
how the knowledge was won , how conflicting theories were resolved, how technological advances made a difference, and so forth ,
anchors
modern
neuroscience
and
renders
it
more
accessible
. The
see
how
even
our
most
secure
convictions
can
turn
out
to
be
come
to go from
to be where
we are is essential
if we are to know
where
here .
What follows is a very brief glimpse into some high points in the
history of neurophysiology . By necessity, the historical sketch is very
selective, and in the main I have followed the particular thread that
has led to an understanding
nervous systems are and a little about their modus operandi . De-
15
ferred until later is that part of the history concerned with the
neurophysiological implementation of psychological functions .
1 .2
Historical
Sketch
By Galen ' s time (200 B.C.) a good deal of the naked -eye anatomy of
the nervous
system
. Galen
was a Greek
anato -
and
that
the
whitish
cords
in
the
muscles
were
somehow
critical . These cords are nerves, and the nerves are really cables containing strands of axon bundles . Galen's hypothesis was that the
nerves transported
muscles and that the muscle then puffed up as the pneuma permeated it , thereby producin ~ movement . In Galen' s conception the
psychic pneuma was breath or air, though as he thought of it , breath
was not merely physical stuff as we now believe it to be, but was
infused with vital spirit . Galen's account was a beginning , though it
uneasily bedded together the mechanistic and the vitalistic , and it
was to persist as orthodoxy until nineteenth -century biologists and
anatomists finally knew enough to replace it .
Descartes (1596- 1650), though sometimes misunderstood on the
matter, had a conception of bodily movement more consistently materialist than Galen. Captivated by the uncanny versatility of clock..
work
mechanisms
and elaborate
water
fountain
systems , Descartes
of
nothing but material bodies and their one peculiarity is that they
are bodies of extreme minuteness and that they move very
quickly like . . . particles of the flame. . . . (1649; in Haldane and
Ross 1911:336)
Clearly I there was nothing very spiritual about his " animal spirits ."
He was especially eager to get a mechanistic account of the reflexes,
for
he saw
such
actions
as instances
in which
16
18
Some Elementary
Neuroscience
reflex. But unfortunately for La Mettrie , the times were far from ready
for such stormy and heretical ideas, and he paid the harsh price of the
iconoclast. He was hounded and reviled by the clergy, banished from
France, and finally exiled even from liberal Holland . Eventually he
was
invited
Voltaire
to
was
the
also
court
of
in residence
Frederick
the
Great
of
Prussia
, where
re-
Nevertheless , the idea that nerves were conduits for animal spirits
gradually lost ground and was put to a particularly telling test by the
great Dutch biologist , Jan Swammerdam (1637- 1680). In one experiment he removed a frog' s leg muscle together with parts of the nerves
attached to it , finding , as others had before him , that the muscle
would contract if the nerve were merely pinched or irritated . He
reasoned
that
if
mere
mechanical
deformation
of
the
nerve
was
claim
that
muscles
move
in virtue
Swammerdam
of an infusion
of pneuma
tested
that
From this he inferred that the muscle changed shape, but that 'I. . . no
matter of sensible or comprehensible bulk flows through the nerves
into the muscles" (Biblia naturae, published posthumously 1738).
Others performed cruder versions of this test on living subjects by
immersing an arm in water , contracting the muscle , and then measur -
and
muscles
roots on the dorsal part of the spinal cord carry sensory information
19
Figure 1.2
Swammerdam
's experimentdesignedto test whether musclevolumeincreasesduring
contraction. At e in the thin tube is a drop of water, which will be causedto rise if the
muscleb increasesin volume when stimulatedmechanically(c) to contract. (Redrawn
from Swammerdam1737- 8.)
20
SomeElementary
Neuroscience
from the periphery to the cord, while the ventral roots carried motor
messagesto the muscles (figure 1.3). This discovery of the separation
of function of the nerves was exceedingly important in establishing
the principle that different functions are executed by different parts of
the nervous system. (As it happened , this was an instance of rediscovery, since unbeknownst to the scientific community , Herophilus
and Erasistratus had recognized this division of labor some two
thousand years before Magendie 's experiments .) The general rule
that dorsal roots carry sensory information inward and ventral roots
carry motor information outward became known as the BellMagendie law , though it appears that Charles Bell (see below ) in fact
had not correctly identified the function of the dorsal roots.
Charles Bell (1774- 1842) pioneered experimental research into the
cause of differences in sensory qualities . By poking himself smartly
on the tongue with a sharp needle, he noticed that for some areas he
could elicit a sensation of pain , but for others he could elicit no pain
whatever , but only a slightly metallic taste. Despite the identity of the
stimulus , the effect was markedly different , and this moved Bell to
believe that the difference was due to the nerves or to the brain and
21
22 SomeElementary
Neuroscience
anatomy and physiology in Berlin , and an impressive number of
famous researchers got their start under his wise and inspiring
tute -
in a nerve
law
was
correct
, and
energy
could
be
transformed
but
neither
23
RECORDING
Stimulated
atA14
- T-.-...!I
Stimulated
at8I
Time
...
.....
Direction
ofDrum
Stimulus
Stimulus
f,
Nerve
fiber f,
/4d ..1
Muscle
(~
ligament
0
Pen
FlECORDING
ORU
~
Figure
1 .4
Schematic
version
impulse
. A
pulls
a pen
when
is
the
nerve
between
apparatus
. This
point
leaves
and
B ) by
a mark
T ( the
stimulated
at B rather
permission
of W . W . Norton
on
the
. This
extra
than
up
B , the
measuring
velocity
used
is set
at point
A . By
impulse
Helmholtz
preparation
is stimulated
at
the
the
- muscle
upward
applied
calculate
of
nerve
muscle
twitch
will
actual
time
is
for
velocity
later
than
if the
, T , he
dividing
muscle
to
a nerve
contracts
, it
showed
twitch
if
:<
- ,y- by~ Henry .
@ 1981
by
that
one
that
stimulus
was
able
d ( the
Psycholo
-
Co . , Inc . Copyright
of
muscle
. Helmholtz
by
the
from
the
difference
obtained
it takes
the
, when
drum
at A ) . ( Reproduced
and
that
a recording
velocity
time
to measure
so
the
nerve
Gleitman
W . W . Norton
to
distance
is
, by-
and
Co . ,
Inc . )
the
touch
the
instant
takes
the
instant
one
time
thoughts
one
was
regards
ing
cessive
but
bodily
and
myself
to
work
regard
as
on
as
or
idea
and
at
its
first
reflection
went
whole
out
business
described
,
admit
appeared
bodily
, a single
I could
' s hand
the
his
own
results
idea
that
' s father
findings
instantaneous
view
The
. Helmholtz
's
, the
the
mental
your
son
touched
reach
shocking
his
your
since
to
rather
regarding
As
was
decided
living
and
I
a
to
me
expression
act
could
star
that
not
that
as
had
only
little
surpris
as
becomes
reconcile
disappeared
suc
24 SomeElementary
Neuroscience
in Abraham 's time should still be visible . (Letter to Hermann von
Helmholtz in Koenigsberger 1906:67)
Another student of Muller 's, Emil du Bois-Reymond (1818- 1896),
was the first to demonstrate (1843) that the nervous effect was in fact
an electrical phenomenon and that a wave of electrical activity passes
down a nerve . It had been well known that nerves could be excited by
" galvanism," but establishing that electricity was the essential feature
of normal nerve function was of great significance and established the
basis for further physiological investigation . Certainly by this time the
idea that a fluid , immaterial or otherwise , is transported in nerves to
cause nervous effects had ceased to be interesting .
The pressing question now concerned the constituents of nerves
and how such constituents were able to produce electrical effects.
Slowly it began to emerge that the basic elements are neurons- cells
with central bodies from which long filaments extend- but this hy pothesis was hard won and was crucially dependent on a variety of
technological discoveries. A number of difficulties obstructed the way
of research here. For one thing , the chromatic aberrations of the early
microscopes meant that artifacts constantly bedeviled observations,
and it was not until the development of the achromatic compound
microscope that it became possible to make reliable observations of
nervous tissue.
Even so, other artifactual problems plagued research, since nervous tissue degenerates unless properly fixated and the differences
between fresh and old preparations are so profound that old preparations are useless. It had to be painfully discovered that water mounted slides were to be avoided because the change in osmotic
pressure changed the cell dramatically . Moreover , as we now know ,
nervous tissue is packed cheek to jowl with cells, some of which are
not neurons at all, but adjunct glial cells. Ingenious stains were eventually found that would highlight select numbers of neurons so they
could be picked out visually from the dense thicket (figures 1.5, 1.6).
Though invaluable , staining was to a troublesome extent an art, and
the resulting preparations did not just emblazon their truths for anyone to read. The observations of the preparations had to be inter preted , and not infrequently there were disputes about what they
truly showed . Finally , it had to be slowly and arduously discovered
that unlike , say, red blood cells, which can be captured in their entirety in the image of the microscope, neurons have long processes
extending well beyond the cell body or " soma."
Histologists , for example Purkyne (1837), saw cell bodies through
the microscope, and on other slides they also saw the long , skinny
25
Figure 1.5
Neurons (Purkinjecells)in the cerebellarcortexof (a) the frog, (b) the alligator, (c) the
pigeon, and (d) the cat. Stained by the Golgi method. (From Llinsand Hillman (1969).
In Neurobiology
ofcerebellar
evolutionanddevelopment,
ed. R. Llins.Chicago:The American Medical Association.)
26
Figure 1.6
Photomicrographof neurons in a cross section of the visual cortexof the mink. The
stain used is cresyl violet (Nissi stain), which stains the cell bodies of all neurons. The
cortex shown here is about 1.2 mm thick, and its six distinct layers can also be seen.
(CourtesyS. McConnelland S. LeVay.)
27
" fibers" that we now call axonsibut it was not until Remak ventured
the opinion in 1838that axons might be extensions of cell bodies that
anyone suspected that neurons were radically different in shape from
blood cells. As staining and fixation techniques improved , this fact
finally became indisputable (Hannover 1840). Using a carmine stain
and new fixation techniques, Deiters (1865) saw for the first time a
luxuriant arborization of very thin and delicate dendrites extending
from the cell body , and on the basis of observed structural differences
he drew a distinction between " protoplasmic processes" (axons) and
" nervous processes" (dendrites ).
These assorted discoveries prompted the question, What are the
connections between neurons ? Connections of soml~kind there must
certainly be, since reflex action clearly required a connection between
ingoing sensory effects and outgoing motor effects, and, following
Magendie, it was known that the sensory nerves were a distinct cable
from the motor nerves.
Camillo Golgi (1843- 1926) took the view that neurons must fuse or
" anastomose" and that in order for the nervous system to do what it
does, it must be one continuous nerve net or " reticulum " (1883). This
became known as the " reticularist theory ." A particularly important
factor in observations was Golgi 's development of a new staining
technique using silver . To understand the fundamental principles of
nervous system organization , it was necessary to know what the
basic structure was, and hence researchers urgently needed a stain
that would allow a single cell to be observed in its entirety . The
staining technique discovered by Golgi had several remarkable prop erties. It selectively stained only a few cells in the tissue examined (110%), and to those selected it did what was wanted : it impregnated
the soma, dendrites , and axons. This meant that for the first time
researchers could observe the fundamental nervC)USunits . The Golgi
technique made it possible to investigate the nervous system in a
powerful new way , and Golgi , by its means and through meticulous
observations, addressed the question of the structural relations between neurons . He concluded , rightly , that dendrites do not anastomose as he thought axons did . He also conjectured, wrongly
however , that dendrites likely served a nutritive rather than a nervous function . His reason for this opinion was that he thought his
stained preparations showed dendrites terminating on blood vessels.
That neurons are independent units and do not fuse to form a
continuous whole was known as the " neuron theory ." To a modern
ear this seems a rather odd label, since reticularists and neuronists
alike believed there were neurons, but differed on whether it was in
the nature of neurons to fuse cytoplasmically or to be separate units .
28
SomeElementary
Neuroscience
However , the word " neuron " was adopted by Waldeyer in his 1891
review of the controversy , and he used it to mean " independent
cell." l Until Waldeyer ' s review , a variety of other expressions were
used to denote what we now call neurons , and indeed the nomenclature was chaotic. This was of course a reflection of the fact that the
nature of the anatomy of nervous tissue was just beginning to be
understood .
The axon anastomoses hypothesized by the reticularists proved
exasperatingly elusive, though considering how tiny is the gap between an axon terminal and the abutting cell body or dendrite , it is
not surprising that some (for example, Held (1897)) thought they had
observed terminals fusing with somas. Golgi staining is a subtle and
rather tricky technique , even now . For one thing , considerable skill is
required to know when the staining is still incomplete , inasmuch as
the stain has not yet made its way to the far-flung ends of the
neuronal processes, and when the staining is past completi9n , inasmuch as the stain begins to impregnate neighboring glial cells. Moreover, not -a little inference and conjecture goes into drawings made
from Golgi preparations , and sometimes things just do not go very
well , especially for the novice .2 Not surprisingly , therefore, the disagreement between the reticularists and the neuronists was not
neatly solvable simply by looking through the microscope at Golgistained preparations . And the controversy was not without heat, for
it concerned a fundamental property of nervous systems, the outcome mattered enormously , and for a long while the evidence was
equivocal. However , by the turn of the century the reticularist hypothesis seemed to have lost considerable ground , and the camp was
composed mainly of diehards .
Experiments with neuronal growth and degeneration proved to be
singularly revealing of the independence of neurons, one from the
other . Wilhelm His (1888) showed in a series of experiments that
foetal neurons definitely start out as independent entities and then
proceed to extend their axonal and dendritic processes. There seemed
no evidence that they subsequently fused . In the mirror image of
His's tests, Forel (1887) found that when a cell body is damaged, only
the axon attached to it degenerates, and conversely, that when an
axon
is damaged, only its cell body shows the typical degenerative
.
sIgns.
Moreover , it was known (Kuhne 1862) that at the neuromuscular
junction axons can be found in special pitted areas of the muscle
fibers, but they do not actually penetrate the muscle membrane . This
was important becauseit meant that axons could transmit their effects
to the muscles, making them contract, without making direct contact
Sys terns
29
came
very
close
to
the
membranes
of
other
cells
but
did
not
net , the
, the
old , vitalistic
theories
would
be
disinterred
and
revived to account for neuronal interaction . As Golgi saw it , the coordinated nature of sensory-guided movement implied that the nerves
were part of a system, and this counted against individual action of
nerve cells . As he remarked in his speech accepting the 1906 Nobel
Prize
for
medicine
I cannot
abandon
the
idea
of a unitarian
action
of the
nervous
figure out should neurons be distinct units is, alas, the discouraging
truth , though the gloomy expectation that mystical forces and substances
as the
would
be invoked
communication
has
between
not
been
cells
borne
is concerned
out , at least
not
so far
tions and despite the purple cast the controversy had sometimes
taken, Ramon y Cajal and Golgi were jointly awarded the Nobel Prize
for physiology and medicine in 1906. Though convinced that neurons
30
SomeElementary
Neuroscience
if all the
nerve
between
centers
motor
formed
nerves
and
a continuous
sensitive
network
and
sensory
structures of
fact, Sherrington discovered it to be much longer - about 100 mil liseconds. Accordingly , he inferred that conduction along nerve fibers
was not the only mode of signal transmission and that the signal must
be
transmitted
across
a gap
between
sensory
neurons
and
motor
came
to be known
as " synapses
."
, researchers
could
observe
the
cell
membranes
and
trace
their
"
came
framework
to
be
for research
established
, in
on neurons
was widely
In this historical
sketch I have touched
in early neuroscience
and have gradually
out research
on the neuron
the
efforts
that
espoused
the
basic
. Nonetheless
sense
33
Systems
, it should
be emphasized
. From
neurologists
(lesions
studying
resulting
the behavioral
from
strokes
effects
or gunshot
of le -
wounds
extremely
important
, did
not
yield
the correlative
" neuronal
ensemble paradigm , " which might be expected to specify
the fundamentals
of what counted as a neuronal ensemble and how it
functioned
concerning
hundred
of neuroscience
years stretching
from
will be briefly
the research
discussed
of Helmholtz
to the introduction
of the electron microscope
into the laboratory
elements for the blossoming
of the science of nervous systems
, the
were
Selected Readings
Brazier, Mary A . B. (1984). A history of neurophysiologyin the 17th and 18thcelzturies.New
York : Raven.
Clarke, Edwin , and C. D . O'Malley (1968). The human brain alzdspinal cord: A historical
study illustrated by zvritings from antiquity to the tzventiethcentury. Berkeley and Los
Angeles: University of California Press.
34
SomeElementary
Neuroscience
Chapter 2
Modern Theory of Neurons
, 1985
2 .1
In trod uction
If we
are to understand
how
the
mind
- brain
works
, it is essential
that
, on
the
number
of connections
between
neurons
, and , per -
whereas tasks of much lesser complexity will take a huge conventional computer days. This immediately implies that however the
brain accomplishes perceptual recognition , it cannot be by millions of
steps arranged in sequence. There is simply not enough time . (This
will be discussed in more detail in chapter 10. Seealso Feldman 1985.)
It is also worth dwelling on the fact that neurons are plastic, that
their informationally relevant parts grow and shrink , that they are
dynamic . Nor is their plasticity a nuisance or an ignorable nicety; it
appears to be essential to their functioning as information -processing
units . Again , as we search for models and theories to understand .the
nature of cognitive abilities , this fact will constrain our theorizing .
36
number
of neurons
and
the
number
of connections
may
be theo -
to none
in the
neuron
itself
, may
well
find
that
the
model
must
postulate many more units than the nervous system has. The number
of neurons and their finite if large number of connections also restrict
the range of possible models (Feldman and Ballard 1982) .
Finally , it is useful to know that neurons and their modus operandi
are essentially the same in all nervous systems - our neurons and the
are
differences
between
vertebrates
and
invertebrates
, but
What matters here is not that this humbling thought pricks our
eminently prickable vanity , but that it reminds us that we, in all our
cognitive glory , evolved, and that our capacities, marvelous as they
are , cannot
be
a bolt
from
the
blue
. Which
means
that
models
for
human cognition are inadequate if they imply a thoroughgoing discontinuity with animal cognition . It is also a reminder that if we want
to understand the nature of the information processing that underlies
, about
their
connections
to other
neurons
, and
about
a methodological
how
con -
ber, to see and problem solve, to care for the young and recognize
danger , it is essential to understand
level
of the
basic
elements
that
make
the
machine
and
at the
level
The human brain weighs about three pounds and has a volume of
about three pints . It contains some 1012neurons, or perhaps as many
as 1014
; the count is only an estimate. When the body is resti~g, the
nervous system consumes about 20 percent of the body ' s oxygen
supply , which is the lion ' s share , considering that the brain accounts
38
SomeElementary
Neuroscience
for only about 2 percent of the body ' s mass and that skeletal muscles ,
the kidneys , the heart , the liver , and so on , also demand oxygen . The
central nervous system (CNS ) consists of the brain and spinal cord ;
the peripheral nervous system (PNS ) consists of all the nervous struc tures external to the brain and spinal cord , such as the fibers innervat ing the muscles and the sensory receptors in the skin . The retina is
considered part of the CNS (figure 2.2) .
Neurons
Neurons are the basic nervous elements and are differentiated
into a
cell body , or soma, and processesl (projections ) extending out from the
soma . The soma is the vital center of the cell , containing the nucleus
and RNA , and it has structures that manufacture protein , much of
which is shipped down the axon by a complex system of axonal
transport . Processes are usually distinguished as axons or dendrites,
but not all neurons have both . Axons are the principal output ap paratus , and dendrites principally receive and integrate signals . Some
sensory neurons in the skin have only an axon , and some neurons in
the olfactory lobe have only dendrites . A single axon generally pro trudes from the soma , and commonly it will branch extensively to ward its end . In contrast , a dense arborization of dendrites often
extends from the soma (figure 2.3) . (See also figure 1.5.) In many
types of neurons the dendrites are covered with stubby branchlets
called spines that serve as the dominant points of contact with other
neurons .
Neurons vary in size , but even the largest is exceedingly small . In
the human nervous system , dendrites may be about 0.5 microns in
diameter , and the soma of a motor neuron is about 20- 70 microns
wide . The largest axons are about 20 microns across , but they are
long - some as long as the spinal cord . There is considerable variation
between different types of neurons , with some showing fairly obvi ous specializations suited to their function . The squid
was discovered
to have motor neurons with relatively large axons (roughly one mil limeter in diameter ) . Given its size , the giant axon of the squid could
be impaled quite easily by recording and stimulating electrodes , allowing the electrochemical properties of axons to be investigated
(Hodgkin and Huxley 1952). (These properties will be discussed in
section 2.3.)
At birth , the primate nervous system has virtually all the neurons it
will ever have . The only known exception is the olfactory system , in
which neurons are continuously induced . Growth of axons and den drites , as well as of the spines on dendrites , is prolific , especially in
the first few years of life . In the midst of this luxuriant growth , how -
40
42
Figure2.4
Types of neurons. The human cerebellumhas over 1010cells but.only five neuronal
types. Eachtype has its characteristicshape, branchingpattern, connectivitypattern,
and position. Seefigures 2.1 and 3.1 for the position of the cerebellumin relation to
other brain divisions. (FromKuffler, Nicholls, and Martin (1984
). FromNeuronto Brain.
2nd ed. Sunderland, Mass.: Sinauer.)
and in some cases axons merely fit into a groove of a neighboring glial
cell . Some neuroglia function as fences (astrocytes ) and as filters
(ependymal cells ) in isolating neurons from blood but not from their
special nutrient bath . Yet others , the microglia , function as phago cytes or scavengers , cleaning up dead neurons and assorted detritus .
The operation of neurons is so dazzling that glial cells tend not to get
their share of the limelight . Nevertheless , outnumbering neurons by
about ten to one , they are crucial to the proper functioning of the
nervous system , though research is only beginning to reveal just how
many tasks they are relied upon to perform . Certainly degeneration
43
Figure
. 5
Diagram
of
layers
( b
around
myelinated
glial
the
glia
Where
Part
of
the
an
1967
the
is
sheath
of
Dembitzer
myelin
myelin
Diagram
myelin
are
in
of
is
figure
( c
of
the
there
lighter
bushes
the
example
up
( a
is
axon
away
to
show
the
completely
segment
cells
and
and
inner
rolled
an
unrolled
up
glial
cell
. )
Schwann
sheaths
Multiple
cut
shown
is
oligodendrocytes
devastating
sclerosis
is
to
one
proper
such
sen
demyelinating
pears
and
and
control
disease
vous
for
make
forms
axon
Hirano
sorimotor
It
that
of
from
that
axon
cell
segment
Modified
of
the
tracts
color
dendrites
presence
gray
of
matter
tissue
this
of
than
for
color
axons
encased
where
which
have
myelin
that
only
axons
difference
in
there
are
distinctly
is
of
grayish
makes
are
myelin
clumps
the
ap
In
with
pinkish
hue
white
section
the
their
between
visible
tissue
and
or
difference
myelinated
easily
the
somas
of
naked
ner
eye
Receptors
Receptors hold a special fascination , perhaps because it is the range of
stimuli to which receptors are sensitive that limits the kinds of things
we sense in the world . Receptors are the interface between world and
brain , and our conception of what the universe is like and what we
45
Figure 2.7
Asectionofthehumanbrainat20degreesfromthespecified
plane.Thecerebral
cortex
showsas the grayrind on the outer surface,followingthe foldsof tissue.The cerebellar
cortexis also visible,as a rind followingthe very deep folds of the cerebellarwhite
matter.
Thecorpuscallosum
consists
ofmyelinated
nervefibers,andsoappears
white.
Thethalamus
contains
a largeconsolidation
of cellbodiesandappearsgray.(From
MatsuiandHirano(1978).
AnAtlasoftheHuman
BrainforComputerized
Tomography.
Copyright
' Igaku-Sh0in
Tokyo/New
York.)
46
Tuberous
organ
Ampullary
organ
(electroreceptor
47
(electroreceptor
WATER
Superficial
neuromast
(mechanoreceptor )
Figure 2.8
Diagram of two different electroreceptors and a mechanoreceptor found in the lateral
line organs of fish . (Modified from Dijkgraaf 1967and Szabo 1974.)
48
types, and (4) voltage-sensitivechangesin single ion channels that transiently open the gates in the channels to permit ions to cross the cell
membrane
The cell membrane is a remarkable sort of sheet, dividing cytoplasm on the inside of the cell from the extracellular fluid on the
outside . The membrane is nonuniformly dotted with tiny pores, specialized to control passage only of certain items . Both the intracellular
and
the
extracellular
fluids
contain
ions , which
are
molecules
or
atoms that have gained or lost electrons and consequently are negatively or positively charged. The plot of the basic electrochemical
story depends on two general classes of ions: large negatively
charged organic ions concentrated inside the cell, and inorganic ions
with systematically changeable concentration profiles inside and outside
the
cell .
The large organic ions inside the cell cannot pass through the membrane, and their net charge is negative. Consequently , this affects the
distribution of ions to which the membrane is permeable, since positively charged ions will tend to congregate inside the cell to balance
the negative charge . The inorganic ions that figure in the story are
potassium (K + ), sodium (Na + ), calcium (Ca + + ), and chloride (CI - ).
the
passage
of Na + and
Ca + + . Thus
, K + concentrates
inside
the
the cell is stimulated , for example by an electric current or by a particular chemical, there is a change in membrane permeability to Na +
and Ca + + . The principal instruments of this change reside in the
structure of the single channel.
49
Na+CIIon
K+A -
K+
Na+
K+A-
Na+CI-
Figure 2.9
Schematic diagram of a neuron soma, showing the internal concentration of inorganic
ions A - and K +, and the external concentration of NA + and CI . The sodium potassium pump in the membrane ejects Na + and hauls in K +. (From Shepherd (1983).
Neurobiology. New York : Oxford University Press.)
What accounts for the voltage drop across the membrane ? Essen tially , the organic anions together with the fact that among cations ,
only K + can cross the membrane to the cell ' s interior . Because the K +
moves inward from areas of low K + concentration to areas of high K +
concentration , it is said to move up its concentration gradient , and it
does so because of the anion attraction inside . It therefore moves
down its electrical gradient . At some point equilibrium between the
two forces is achieved , in the sense that there is no net movement of
K + across the membrane , and the electrical force required to keep K +
50
SomeElementary
Neuroscience
-
--
+
+
+
-
+
-
+ - +
+ -
+ -
+ -
+
-
+
-
+- -
- + + - +
+
+
-
+
+
-
+
+
+
-
+
+
+
+ +
+ +
+ +
+
+
+
-
+
+
+
+ +
+ - +
+
+
+
+
+
-
+ -
+ -
+ -
+
+
+
+
+
-
+ +
+ - +
+ -
+ -
+ - +
+ -
+ -
+ +
+
-
+ +
+ +
+ - + - + - +
+
+ -
+ +
+ - + - +
- + - + - + - + + - + - +
+
-
+ - + - + - + - + - +
+
-
+ - +
+
-
+ - + + - +
+ -
+ - +
- + + - +
- + + - +
- + -
-- +
+
-
+ - + +
-- + - +
+ + -- +
+ + + - + - +
+ + + + + - +
- +
+
+ - + - + +
+ - + - + - + +
+ -
+ -
+
-
+ - - + - + - +
+
-
+ -
-- + -
+ -
+
+
+ -
+ -
+ -
+ - + + + - +
- +
+ -
+ - +
- + + + - + - + - + +
+ - + - + - + - +
+ + + - + - + - + + - + - + - + - + - +
+
+
+
+ + + + - +
- + - + - + + - + - + - +
- + - + - + + - + - + - +
+ -
+ - +
- + + - +
- + + - +
- + -
- -
+ -
- + + - +
- + + - +
+ + - +
+
+
+
-
- + - +
+ - + - + - +
+ - + - + - +
+ -
Figure 2.10
Schematic cross section of a neuron process showing the concentration of negative
charges along the inside of the membrane and positive charges along the outside .
(Reprinted with permission of the publisher from Koester (1981). Ch . 3 of Principlesof
Neural Science
, ed. E. R. Kandel and J. H . Schwartz, pp . 27- 35. Copyright ~ 1981 by
Elsevier Science Publishing Co., Inc .)
at its concentration
gradient can be calculated . This calculation
yields
the electrical potential for K + across the membrane . For example , in
some neurons the equilibrium
potential
for K + (no net movement
of
K + ) is - 70 millivolts
(mv ) . The electromotive
force is the force tend ing to equalize
is a measure
in
charged
with
respect
to the outside
(its resting
potential
).
51
Figure2.11
Idealizedexperimentfor measuringthe potentialdifferenceacrossthe cell membrane.
The electrodeis a fine glass capillary with a tip no more than 0.1 micrometerin
diameterfilled with a salinesolution.
STIMULUS
ELECTRODE
STIMUL
ELEC
t KINJE
t
CELL
BASPURK
CELL
. .
. .
Modern Theory of Neurons
53
I
I
Figure2.13
The somaof a Purkinjecell in the vicinity of stimulationis briefly depolarized(upward
deflection), the degreeof depolarizationbeing a function of the stimulusintensity. A
Purkinje cell outside that preferredarea receiv
'es inhibitory input from the activated
axon of a basketcell and is transientlyhyperpolarized(downward deflection). (Courtesy R. Llinas.)
54
L_
__
~EPSP
__---~L__y__
~
Graded
55
Trigger:
all.or-none
spi
Figure
diagram
impulse
showing
initiation
potentials
the
EPSPs
axon
( From
cell
or
spi
none
ke
spi
to
ke
next
ce
II
iated
travels
1967
motor
neuron
bodies
synaptic
and
on
and
inhibitory
hillock
concentration
of
respond
potentials
) ;
down
of
the
events
with
( IPSPs
undiminished
Foundations
gradients
channel
has
open
for
channel
mean
time
PS
responsible
graded
the
~1
potential
axon
~ / cholog
~/
for
excitator
action
the
Physiological
of
because
the
special
Suppose
by
Na
the
cell
yet
results
in
the
more
Na
generating
is
IPSPs
are
New
5 ~/
triggered
not
York
l1aptiL
shown
Harper
and
is
large
efflux
of
influx
of
Na
ions
In
absolute
studied
sensitive
cases
the
and
1980
total
cross
substantial
in
is
When
current
be
this
each
picoamps
Neher
can
the
move
itself
can
change
and
cause
in
allow
the
positive
net
influx
terms
Na
then
Na
If
of
feedback
number
to
the
Thus
self
depolarizing
is
greater
in
of
channels
enter
changes
initial
axon
inward
-t -
influx
the
Na
the
This
induces
results
the
Na
more
further
cell
additional
which
produced
of
the
even
further
is
membrane
inside
allowing
to
the
the
therefore
ions
transient
that
thereby
effect
In
and
depolarize
about
Sigworth
Na
channels
enough
membrane
explosive
15
millisecond
voltage
as
depolarizing
current
membrane
signals
mv
of
msec
effect
10
great
are
about
axon
in
amplifying
current
sufficiently
incoming
hillock
membrane
channels
self
figure
current
of
is
patch
channel
mean
density
ing
in
11
of
. )
and
in
location
Dendrites
or
Thompson
Row
ini
ion
. 14
Summary
in
ke
Conducted
( conduct
ions
than
sudden
crossing
the
large
the
56
AT REST
CLOSED
DEPOLARIZED
OPEN
Out
DEPOLARIZED
INACTIV
ATED
I
SeJectivi ty
filter
In
I J
Inactivation
gate
Activation
gate
.
.
2 nm
'
.
H2ONa"
Figure2.15
Voltage-sensitivesodiumchannel, drawn schematicallyto scaleaccordingto biochemical, electronmicroscopic
, and electrophysiological
information. Ionic selectivityis provided by a constriction lined with negative chargesnear the outer surfaceof the
membrane.Theactivationgatenearthe inner surfaceopensin associationwith translocation of negativechargesacrossthe membranefrom out to in. The inactivationgate
blocks the inner mouth of the channeland preventsclosing of the activationgate.
Watermoleculeand hydratedsodiumion aredrawn to scalefor comparison.(Modified
from Kuffler, Nicholls, and Martin (1984). FromNeuronto Brain. 2nd ed. Sunderland,
Mass.: Sinauer.)
57
Membrane
depolarization
Increased
Membrane
Increased
sodium
depolarization
potassi u m
conductance
conductance
Repolarization
Sodium
A
SPII
\ 'f
Potassi
influx
( " AlTION
PUI
~ " rIAL
" )
+ 40
um
efflux
~
INT
~~VAL
=6.' MILLISlCONDS
VELOCITY
2 ')
m / sec
tlO
MEMI ) RAN ~
POTFNTIAL
( MILLIVl
- - -
- 70
) LTSJ
- 40
-
RESTING
POTENTIAL
0 \)
- HO
No "
. t
+
', '
No "
. .
+
TT T
A\ ON
-+
"
-+ +
+ ~
-t
-t
~ -
..
,
S
No "
r
CENTIMETERS
J
K+ 4.
-+
l -n
. :
K+ 4.
-
I
10
No "
I
15
I
20
I
25
I
30
C
Figure 2 .16
(A ) Positive feedback effect resulting from above threshold depolarization of the mem brane . (B) Restoration of the membrane ' s resting potential . (C) Propagation of a nerve
impulse along the axon . The large change in potential is initiated by a small influx of
sodium ions , which opens voltage -sensitive sodium channels , changing the potential
further . The membrane ' s resting potential is restored as the sodium channels are inac tivated , and potassium channels open to permit an outflow of potassium ions . This
sequence of events begins at the axon hillock and continues down the length of the
axon .
58 SomeElementary
Neuroscience
duced by the impulse will appear on the screen as a spike; hence,
impulses are also referred to as " spikes" (figure 2.16).
During the brief interval when the membrane is permeable to Na +,
the potential across the membrane at the relevant segment changes
enormously as a consequence of the inward Na + current . This current will spread along the membrane, which will cause depolarization
in the adjacent areas of the membrane, and Na + channels located
there will , in their turn , undergo a conformational change to allow
Na + current , thereby engaging the regenerative process to permit an
infl ux of N a + ions in that region , and so on down the length of the
membrane (figure 2.16). Therefore, the drama in the axon does not
end with the production of a localized spike, for when an action
potential is produced at the hillock , the spreading current depolarizes
the neighboring membrane downstream , which in turn generates an
action potential and consequently depolarizes its downstream neighborhood membrane, and so on .
In this fashion , a wave of depolarization and repolarization travels
from the trigger zone in the axon hillock down the length of the axon.
(It could travel the reverse direction , and can be made to do so in the
laboratory , but in the un tampered neuron it does not .) The signal
does not alter in its journey down the axon, becausethe amplitude of
the spike does not diminish as it travels . As long as there is one spike,
this ensur ~s that the adjacent membrane will be depolarized above its
threshold , which means it will spike, and so on to the fiber end.
During its spiking phase the axon cannot spike again; this is called its
refractory period (figure 2.16). One might think of this by analogy
with a slingshot that cannot fire again immediately but requires an
interval for the sling to be repositioned and to regain a store of energy . TIle single channels have to be reconfigured , Na + has to be
pumped out , and the neuron membrane has to regain its balance of
electric potentials .
In view of the importance of time constraints imposed by the nervous system on modeling, it should be mentioned that the time course
for a spike is some 0.2- 5.0 msec, depending on the neuron , and this
means that there is an upper limit on the spiking frequency of any
given neuron . In humans some neurons can spike 500 times per
second. For purposes of rough calculation, let us say that a spike
takes about 1 msec. Now if a perceptual recognition task takes about
100 msec, then there can be at most 100 information -processing steps
between input and output . Models that require ten thousand or a
million steps are going to be out by several orders of magnitude .
The basic account of the electrophysiology of neurons w ~s worked
out by Hodgkin and Huxley in 1952, but not until recently has the
Modern
Theory of Neurons
59
60
B
---- -
- - -- - -
c
D
IIfIIII1II1
'I'I,I:iIII
I'I:!,'!I;1,:i1
'I:'.:1I"I!', i!:IIII,,JI'/,1"
"I'II
I ~I I I,I,:I
~
50msec
20mV
Figure 2 .17
Sodium and calcium action potentials in (A ) a Purkinje cell . (B) Depolarization
of the
dendrite produced long -duration calcium action potentials , and (E) depolarization
of
the soma produced high -frequency sodium action potentials , interrupted periodically
by a calcium action potential that was followed by a transient hyperpolarization
. (C)
and (D ) represent the effects of the passive spread of depolarization
from the soma .
(From Llinas and Sugimori 1980 .)
producing
bined
special
use
neurons
show
externally
or
a low
induced
Ca + + ) , and
currents
2 . 19 ) .
and
Why
must
and
patterns
a train
the
sending
rate
this
uses
by
hundred
in
times
be lavish
. Evolution
elin . The
strategy
cover
of
firing
energy
now
. If a neuron
routinely
space
is that
of a second
, its energy
upon
if glial
depolarizing
one
cells
by
be evident
2 . 18 ,
. Neurons
essential
to receiving
handles
a thousand
or
so , and
- saving
ensheath
current
of Na +
depolarizing
( figures
consumption
energy
com -
without
leakage
currents
will
the
. Frequently
( spiking
is increased
gradients
stumbled
, then
spiking
by inward
ionic
a second
through
currents
hyperpolarizing
the
the
here
impulses
, perhaps
so much
information
of
depolarizing
spontaneous
rate
maintain
signals
an insulating
of
base
decreased
brain
depolarizing
and
depolarization
continuously
several
in
of hyperpolarizing
from
if it spikes
will
device
have
to
in my -
the
axon
to form
one
action
poten
61
Figure 2.18
Interactions of excitatory and inhibitory synaptic potentials (EPSPsand IPSPs) in an
otherwise silent cell. Each of the synaptic potentials illustrated here is usually produced
by the synchronous action of many presynaptic neurons . (Reprinted \\,ith permission
of the publisher from Kandel (1981a). Ch . 7 of Principles of Neural Science
, ed. E. R.
Kandel and J. H . Schwartz, pp . 63- 80. Copyright @ 1981by Elsevier SciencePublishing
Co ., Inc .)
Figure2.19
Sculptingrole of inhibition, shown here to producechangesin the firing pattern of a
spontaneouslyactiveneuron. (Reprintedwith permissionof the publisherfrom Kandel
(1981a
). Ch. 7 of Principles
afNeuralScience
, ed. E. R. KandelandJ. H. Schwartz, pp. 6380. Copyright @1981by ElsevierSciencePublishingCo., Inc.)
tial will travel further down the axon and so the energy-intensive
action potential need occur only at wider intervals . This is called
saltatory conduction, because the spikes, as it were, jump down the
axon in long strides (figure 2.14).
Rolled-up Schwann cells are strung along peripheral fibers like sausageson a string , and the action potentials occur only at the exposed
membrane between the Schwann cells, the " nodes of Ranvier." A
large, well -myelinated axon in a human motor neuron can conduct an
impulse up to 130 meters per second, whereas an unmyelinated fiber
is much slower , sending impulses at only about 0.5 meters per sec-
62
SomeElementary
Neuroscience
Neuronal Integration
There are two fundamental types of connection between neurons :
electrical synapses and chemical synapses. Electrical synapses are of
two types: (1) those generating field potentials, in which sending and
receiving neurons are so closely positioned that current flow in one
induces field changes in its neighbor , and (2) gap junctions, which
consist of supremely thin protein tubes connecting the axon of one
neuron to the dendrite or axon of another (figure 1.8). The tubes are
so narrow as to permit the transfer of only very small ions such as
Na + or K +, and it is via the transfer of these ions that signals are
transmitted from one neuron to the next .
Electrical synapses were for some time believed to be unique to
primitive nervous systems, and though demonstrating their existence
in the mammalian CNS is extremely difficult , research in the past ten
years has shown electrical coupling in cells in the hippocampus and
cells in the inferior olive that project to the cerebellum (Llinas , Baker,
and Sotelo 1974). The intriguing question now is whether electrically
coupled cells have a special functional significance in nervous systems.
The leading hypothesis focuses on the major difference between
chemically coupled cells and electrically coupled cells, namely that
the absence of synaptic delay (see below ) ill electrically coupled cells
means that they can fire synchronously . Such synchronicity , along
with positive feedback, appears to have an important role in generating rhythmic patterns typical of various CNS structures (Bower and
Llinas 1983, MacVicar and Dudek 1980). In the caseof the cells in the
inferior olive , the electrical coupling may serve to establish synchronous firing of bands of Purkinje cells in the cerebellum. Since Purkinje cells are known to be crucial in subserving sensorimotor
coordination , this general line of research has suggested that the
synchronizing of rhythmic patterns in sets of neuronS may embody a
fundamental principle of neuronal organization underlying sensorimotor coordination (Llinas 1984b, 1984c).
Chemical synapses (figures 1.7, 2.20) have been most intensively
studied in the giant synapse of the squid, and at the synaptic terminal
it is Ca + + ions and Ca + + channels that play the crucial role (Llinas
1982). When a depolarizing wave reaches an end bulb of an axon, it
opens voltage-sensitive Ca + + channels. Ca + + rushes into the cell
and causes little vesicles containing neurotransmitter substance to
fuse with the outer membrane at specialized zones (Heuser and Reese
64
0.......
Figure 2 .21
fuses with end bulb membrane . (After Heuser and Reese 1979.)
substance
is
into
the
extracellular
space
that separates the axon from the adjacent neuronal process . Some of
the neurotransmitter diffuses across the synaptic cleft and binds itself
to specialized sites on the receiving cell - the postsynaptic membrane
the synaptic cleft . The actual time of the synaptic delay is about one
millisecond , which is remarkably short considering the complex
molecular scenario (Llinas 1982) .
66
67
Presynaptic neuron
Control
Posttetanic potentiation
Tetanus
mV;::~--.t"'-----."'i
+ 40
- 60
Time
- +-
Postsynaptic neuron
mV
2 .5
1.
Time - . . . .
Figure
2 .23
Post - tetanic
synaptic
speed
tic
potential
target
: simultaneous
of this experiment
potential
neuron
aptic
as a simple
is stimulated
tetanus
the
neuron
is returned
stimulated
PSP increases
in size . After
continues
to be
facilitated
( Reprinted
with
permission
Neural
~ ~~ ~E !JJW
~
Science
( Kandel
1979
released
a>t!ds J~1deuAsaJd
apn1!ldw ~
) are
hare
detail
the
to
venom
in
acetylcholine
muscles
common
Kandel
and postsynap
the
and
in
1 mv . The presyn
of tetanus
some
Kandel
presynaptic
of l / sec ; however
, the
this
presynaptic
for
several
hours
( 1981b ) . Ch . 8 of Principles
, pp . 81 - 90 . Copyright
.
of
@ 1981 by
binding
food
. This
produce
, therefore
poisoning
be
changes
in
neurotransmitter
sensitization
be
in
discussed
as
( the
to
modification
that
used
curare
binding
to
to
organisms
to
breakdown
muscles
and
will
synapses
by
, like
from
that
of
in
the
more
.)
by
. The
shown
amount
example
chemical
) , works
) have
the
habituation
( This
neurotoxins
of
1984
in
producing
- bungarotoxin
cause
seconds
period
a PSP of about
J. H . Schwartz
exploited
produce
resisting
control
at a higher
from
changes
subsection
of
Indians
and
and
been
alpha
Amazon
(a
next
has
capacity
tract
minutes
Californica
vulnerability
chemicals
cles
many
hence
central
Aplysia
in
The
( and
neuron
a long
Co . , Inc .)
, Hawkins
Ca + + current
several
of firing
and
the
seconds
of the publisher
Publishing
over
rate
Science , ed . E . R . Kandel
Elsevier
sea
for
occur
produces
for several
to its control
a presynaptic
that
line . ) During
is then
from
events
appears
neuron
recordings
to show
special
arrow
receptor
the
depolarizing
) also
paralyzed
produces
assorted
evolved
weapons
prevents
, are
by
have
the
. The
poison
sites
snake
used
on
by
mus
con
neurotransmitter
current
. Botulinus
paralysis
to
toxin
,
but
68
SomeElementary
Neuroscience
achieves this by preventing the release of acetylcholine from the presynaptic membrane . In this casereceptor sites are available, but there
is no acetylcholine to bind to them , so once again paralysis is the
result . Other neurotoxins , such as black widow spider venom , in crease the discharge from the presynaptic membrane, causing the
muscle cells to be excessively stimulated , resulting in rigidity and
tremor .
Although the human nervous system has not evolved venom
pouches or poison sacs, we have learned how to synthesize certain
neurochemicals in the laboratory and how to use them to intervene
directly in the neurotransmitting affairs of neurons . Neurophar macology is the study of chemicals that affect neurons, and in recent
times it has become an area of intense research as scientists try to
discover effective treatment for diseases of the nervous system. Perhaps because of its immediacy to clinical concerns, neuropharmacol ogy has become a glamor discipline within the wider domain of
neuroscience. Three discoveries in particular have propelled it into
public attention .
First was the discovery in the 19505that certain drugs dramatically
curtailed psychotic symptoms in many patients, enabling them, if not
to lead completely normal lives, at least to live outside the asylum
walls . In the short run these drugs calm violent and wildly excitable
patients, and in the long run they abolish hallucinations and ameliorate disorders of thought . Since schizophrenia is a devastating and
widespread mental disease, finding even a consistently palliative
drug has had profound social significance. Such findings naturally
engendered the hope that knowledge of how the drugs worked
would lead to knowledge of the disease and its etiology , and thereby
to knowledge of prevention .
Second were discoveries, also in the 1950s, that led to the treatment
of Parkinson's disease, also known as the shaking palsy . Parkinson's
disease is characterized by muscular rigidity , tremor , and akinesia (a
diminished ability to make voluntary movements). It was found in
autopsies that the brains of patients with Parkinson' s disease had
abnormally low levels of the neurotransmitter dopamine and that one
motor area rich in dopamine -producing neurons (the substantia
nigra) had degenerated. This suggested that dopamine deficiency
was the root cause of the disease and that the motor dysfunctions
could be alleviated by giving the patients the drug L-dopa (which
converts to dopamine in the brain ). The idea was tried , with considerable though not unalloyed success, and L-dopa is now the drug of
choice for reducing the effects of Parkinson's disease. This too was an
important discovery , for Parkinson 's disease afflicts large numbers of
69
in the
treatment
of Parkinson
1982.)
Third was the revelation
' s disease
, see Larsen
and
CaIne
that
there
are
at
least
five , three
of
which
were
classed
as
" endorphins " and two as " enkephalins ." Though this discovery suggested a practical application in the relief of pain and of mood disor ders, it also raised many questions . What are the opiates doing in the
brain in the first place? Will we find endogenous tranquilizers and
endogenous antidepressants?3 Are certain diseases of the mind
caused by imbalances in these chemicals ? Can I be addicted to my
own
chemicals
Investigation
of the neurochemicals
it shows
us that
chemical
events
at the
cellular
level
can have
cal theory to be irreducible to neurobiological theory . (Seechapters 79.) Not that neuropharmacology can now yield anything like a
decisive demonstration of the falsity of these views , but it can under mine certain favored theses about how very different and separate are
conviction
mound of biological stuff hidden under the skull . It can help to shift
the burden of proof to those who deny that there can be a science of
the mind . Therefore , after a few simple illustrations
of neurons as
they participate in networks, I shall dwell a bit further on neurophar macological considerations .
Some Simple Wiring Diagrams
To understand
what
the
brain
does , it is necessary
to understand
the
intervening network is typically exceedingly complex, studying examples in which the neuronal connections between sensory input
and behavioral output are very simple has been an important
step in
70
seeing how input -output effects are achieved and in developing models of the intervening
information
To amplify the earlier mention of the cellular basis of simple habitu ation and sensitization , I shall illustrate the revolutionary discoveries
in the neurobiology of behavior with the neuronal circuits in Aplysia
(figure 2.24) that mediate gill withdrawal following stimulation of the
siphon , habituation to a gentle stimulus , and sensitization after a
painful
stimulus
stimulus
In Aplysia the circuit leading from the siphon 's sensory periphery to
the motor periphery of the gill is very simple , as can be seen in the
schematic wiring diagram in figure 2.25. It shows one of the 24 sensory neurons innervating
the motor neurons innervating the gill . (Only one of the six gill motor
neurons is shown .) The connection
sensory
and
motor
neurons
with a gentle squirt of sea water , and after a few trials the
junction , with the result that the ~ otor neuron was less depolarized
and hence caused smaller contractions in the gill muscle.
In sensitization roughly the opposite happens; there is an increase
in Ca + + current
and
hence
an increase
in the
volume
of neurotrans
mitter released into the synaptic cleft . However , this effect requires
(presynaptic
facilitation )
(figure 2.25). When the tail is given a noxious stimulus , the facilitat ory interneuron
Modern
Mantle
71
Theory of Neurons
shelf
Eye
.
"
"
,
,
,
,
J
,J
J
Tail
Siphon
""' "
"'
,
'.
"
'
'"
. ...
Foot
'.,
', ',
" '
'
, .
'
. .
' ,'
"','
,
:.': ,".
,' ,
.". ,. "..-
Eye
Figure 2.24
Top view and side view of the sea hare, Aplysia Californica. When the mantle shelf is
stimulated , the gill contracts.
72
SomeElementary
Neuroscience
Tail
Siphon
Figure 2.25
Partial neuronal circuit for the Aplysia gill and siphon withdrawal
reflex and its
modification . Mechanosensory
neurons (S.N .) from the siphon make direct excitatory
synaptic connections onto gill and siphon motor neurons . Tail sensory neurons excite
facilitator interneurons , which produce presynaptic facilitation of the siphon sensory
neurons . (From Hawkins and Kandel (1984). In Neurobiology of Learning and Memory , ed .
G . Lynch , J. L . McGaugh , and N . M . Weinberger . New York : Guilford .)
stimulus
guity
to which
it is sensitized
of CS and
sensory
UCS
neurons
is not
yet
Each
on later
understood
of these
there
might
of the
events
various
nation
of the
they
argue
extinction
nations
that
hypothesis
To further
model
for
found
branches
the
temporal
conti
of Ca + + current
exactly
forms
habituation
how
this
in the
happens
cellular
of
plasticity
event
employs
- types
sequences
of
a distinct
have
kinds
each
combi
. They
plasticity
learning
reductive
events
of
that
that
2 .26 ) . Possibly
as associative
of cellular
models
suggested
sense
( figure
might
, classi -
set of specified
, " in the
such
generalization
small
( 1984 ) have
" alphabet
plasticity
evaluation
, sensitization
a relatively
Kandel
cellular
of precise
expla present
based
on
.
illustrate
a complex
been
and
of behavioral
stimulus
in terms
experimental
yield
types
to employ
Hawkins
fundamental
, and
that
, though
of plasticity
a general
, other
for
occasions
appears
, and
exist
. It appears
an enhancement
forms
cal conditioning
cellular
causes
contrast
that
how
effect
groups
, consider
of neurons
the
enhancement
sensory
) to interneurons
. In
neurons
that
circuit
have
in
might
be connected
figure
2 .27 , which
every
studied
organism
typically
send
collaterals
an inhibitory
effect
to
is a
it
has
(axon
on the neigh
Habituation
Sensitization
Classical conditioning
73
S
T
~
R
SE
A
i.ATP
J.PHO
c
L
5
'
AM
AM
.Ix.C
AM
S ROTONINREC
PTOR
- - G PROTEIN
- ADENYLATE
CYCLASE
Co
, . BINDING
SITECo
..
::u"
PHOSPHOPRQTEI
N
PHOSPHATASE
S
-KTEIN
+CHANNEl
P~
-PO
~
S K. C-' ANN[ L
;:r;J' [ I~ - PO4
Ca
'.
P-O.O
POST
~ IC
CElJ
. ""'\
Co, .
!QSTSY
!'.APTI
( CEll
.
~ "\
J~ I
Figure 2.26
Cellular mechanisms of habituation , sensitization , and classical conditioning of the
Aplysia gill and siphon withdrawal reflex . (A ) Habituation : Repeated stimulation of a
siphon sensory neuron (the presynaptic cell in the figure ) produces prolonged inactivation of Ca+ + channels in that neuron (represented by the closed gates), leading to a
decreasein Ca + + influx during each action potential and decreasedtransmitter release.
(B) Sensitization : Stimulation of the tail produces prolonged inactivation of K + channels in the siphon sensory neuron through a sequence of steps involving cAMP (cyclic
adenosine monophosphate ) and protein phosphorylation . Closing these channels produces broadening of subsequent action potentials , which in turn produces an increase
in Ca + + influx and increased transmitter release. (C) Classical conditioning : Tail stimu lation produces amplified facilitation of transmitter release from the siphon sensory
neuron if the tail stimulation is preceded by action potentials in the sensory neuron .
This effect may be due to " priming " of the adenyl cyclase by Ca+ + that enters the
sensory neuron during the action potentials , so that the cyclase produces more cAMP
when it is activated by tail stimulation . (From Hawkins and Kandel (1984). In Neurobiology of Learningand Memory, ed. G. Lynch , J. L . McGaugh , and N . M . Weinberger . New
York : Guilford .)
11
1 nl
I .
boring sensory neurons . This is called lateral inhibition , and it appears to be a common arrangement in nervous tissue, being found in
the retina , the skin , the olfactory epithelium , and the gustatory
epithelium . The effect of a lateral inhibition circuit is to enhance the
contrast between highly stimulated neurons and their nonstimulated
neighbors , since the stimulated cells fire at a high rate and the inhibited cells fire at a rate lower than their base rate. Some such arrangemen t in the retina is believed to figure in the perceptual effect
known as J'Mach bands." The effect is easily seen in figure 2.28. The
74
Some Elementary
Neuroscience
o +S
<~ "'}
0S
0.
~ '"
os
'~""y S
~ '~",,:fOS
os
<= 'jO
' "
>. 0
' "
"
-t-1
"
~ O
A
0. .
"' ~
> .."
~
vS
/""'YQ
~.-~"'~~'o~ -~,
"
~
S
'-'"
/~ 0
' " ~~ -_ --:!
o.. ",,;'
o' " S
~= ~YO
~
,
...
"
OS
a.
", '"
" "
0"
C~
' ,
~=
'~Second
o-0
Receptors
Inhibitory
sensory
-order
neurons
Interneurons
f"'\ S
'-'
/ 0
""' Y'"'
"' -+ S
, ;:-0 -
Q.
,,"
0S
o' " S
<~ ~YO
~
OS
O
" " S
yo ~= :~
~ S -r- E
/ 0"
"
"
...
Q.
,, ~ >.
+1
...
" "
a. .
... " ~
. . .
" ,
'-'
'
" ~o ~-~
-~ 7
"
"
'
f"'\ S T E
'-'
("\ S -t- E
/ 0' " ~~ 2 ~ ~
"
"
....,- -
--
"
~ ~
...",; :
"
"
~~-=E
Receptors
"
,
<~ "<~..:~ ~
0
Inhibitory
Second - order
sensory
neurons
Interneurons
"
. .
Physiological
intensity
A
" Brighter
' 4
Stimulus
effect
Figure 2 .27
Lateral inhibition . (A ) A pattern of uniformly gray areas with sharp edges, representing
a visual field seen by an eye. (B) The stimulus plotted as intensity (horizontal ) against
spatial extent (vertical ), emphasizing the uniformity of the physical intensity within
each area. (C) A network of receptors and second-order neurons with reciprocal inhibi tory connections via interneurons (broken lines). Spontaneous activity in the receptors
(S) is augmented by excitation (E) due to light . The network causes the second-order
neurons to show S activity augmented by E and/or reduced by inhibition (I) in single or
double (21) dose. (0 ) The output , equivalent to our sensation, plotted as darker or
lighter than the background due to S. This sort of arrangement probably explains the
Mach bands illusion seen in figure 2.28. (From Bullock , Orkand , and Grinnell (1977).
Introduction to Nervous Systems. W . H . Freeman and Co. Copyright @ 1977.)
75
POSITION
Figure 2.28
Although there is a smooth reduction in light intensity from right to left in the photograph, there appears to be a bright band at Y and a dark band at X. These are Mach
band patterns. (Modified from Cornsweet (1970). Visual Perception. New York: Academic Press.)
76
Mach bands visible in the boundary area between a light region and a
dark region appear as a sharper change in light intensity than actually
obtains.
How contrast enhancement works in the hypothetical circuit can be
observed in figure 2.27. All cells have a base firing rate, cells stimu lated by the light have an increased firing rate, and each inhibitory
neuron decreases the firing rate of its postsynaptic cell by some
specified amount . Consider the cell that is on the boundary of the
increase in stimulus intensity . Excited by the stimulus , it will fire at a
higher than base rate, and it will thel'efore excite an inhibitory inter neuron that synapses on the cells that are its immediate neighbors .
One of its neighboring cells, unaffected by the light stimulus , will
have its base rate decreased by the inhibitory interneuron , while the
neighbor on the other side, despite being excited by the light
stimulus , will fire less vigorously than the boundary cell because it
receives inhibitory stimuli from excited neighbors on both sides.
Since the cells at the boundary of the area receiving the stimulus are
inhibited only by neighbors on one side, they have the highest firing
rate. Although this is a highly simplified and schematic representation of how a network of neurons might interact to yield contrast
enhancement, it illustrates how complex effects can be derived from
appropriately connected simple units ,
These examples of simple circuits are meant to provide only an
introductory illustration of how neurons can be wired up to constitute
an information -processing and a learning network . Nevertheless,
they should help to make clear why neurobiologists have insisted it is
necessary to understand the architectural and physiological proper ties of nervous systems. The reason is that such properties provide
essential clues both to how networks can generate complex representations of the external world and of the internal nervous-system
world and to the nature of the computations performed by large,
interconnected arrays of neurons . In chapter 3 I will say more about
some architectural properties of nervous systems believed to be especially significant in constructing a theory of the nature of neural
network function , In chapter 10 I will discuss some theoretical
developments arising out of, or at least directly inspired by, neuroanatomy and neurophysiology .
ConcludingRemarks
Although the membranes of all body cells are polarized and have the
ability to depolarize and repolarize , neurons are special both because
their single channel configurations permit them to exploit this capacity in a coordinatedand systematicfashion and because they are joined
together
in
tures
of
with
the
network
and
muscular
The
of
principle
the
capacities
of
packed
the
matter
of
From
system
' s
and
motor
in
of
that
improves
2 .4
and
view
of
system
neurotransmitters
not
upward
even
for
close
stances
tainty
that
more
these
extravagant
in
diversity
in
These
of
some
of
others
others
or
brain
,
;
mainly
the
neurons
building
- in
of
increasingly
fancy
.
As
the
for
in
more
world
- repre
ramify
been
predicted
one
for
the
Has
or
is
active
and
it
is
been
the
sub
near
Nature
does
as
Even
prediction
themselves
the
serve
.
neurally
revealing
variety
to
inhibition
scrounge
known
brain
that
chemicals
system
into
the
that
;
,
cer
What
are
wantonly
variety
believed
some
some
are
others
at
These
reflect
synaptic
have
quite
produce
hopelessness
;
,
others
syn
euphoria
some
drinking
produce
and
they
heighten
mem
,
eating
changes
,
of
consistent
facilitate
excessive
junctions
the
action
hallucinations
to
amnestic
of
the
produce
moods
drugs
light
obstruct
can
produce
.
or
may
alter
and
in
,
Some
behavior
is
enhance
moreover
anxiety
;
pain
sexual
fascinating
,
effects
tranquilize
alleviate
it
particular
innate
two
to
forty
nervous
and
despair
tendency
mimic
hallucinations
unseemly
a
the
pressure
and
especially
that
substances
others
some
just
least
in
psychological
produce
arousal
have
coaxed
become
reduce
might
produce
such
drugs
these
merely
certain
distinct
densely
Neurochemicals
the
producing
function
questions
thesis
and
be
the
thing
learning
plasticity
at
doing
In
the
for
evolutionary
excitation
in
will
substances
permits
the
' s
around
that
sensory
through
it
Nature
changing
funda
not
means
for
of
in
marvelous
is
provide
be
same
those
ways
it
time
neurons
however
between
can
same
the
The
of
to
are
dimensions
things
There
sloshing
it
Other
one
coordination
need
adjusted
the
would
follow
remarkable
and
on
sometimes
under
and
simple
nervous
but
the
; the
just
- representation
Neurotransmitters
On
features
essentially
type
work
neurons
sensorimotor
sentation
fea
such
at
components
revolutionary
world
but
will
ganglia
and
essentially
adding
proliferates
competitive
not
does
novel
is
pool
it
that
- representation
- dating
brain
interposes
neurons
world
represent
of
basic
human
are
system
that
same
flatworm
is
capacities
terneuron
ory
brain
the
step
basic
of
enlarging
simple
the
which
circuits
evolutionary
all
in
human
electrical
can
occurrence
complexity
of
conglomeration
about
neurons
deadly
Increased
neurons
principles
that
the
is
.
components
and
is
versatile
adding
flatworm
mental
result
coordinate
movement
endlessly
matter
The
can
fundamental
almost
up
world
77
in
provide
,
the
a
78 SomeElementary
Neuroscience
means for interfering in the nervous system that may help us under stand
what
various
parts
of the brain
do .
The assorted endogenous neurochemicals are not distributed randomly in the brain . First to be considered is Dale's law , which says
that any given neuron synthesizes and releases only one type of
transmitter , though it may have receptor sites fitted for any number
of diverse transmitters . The law seems to hold generally for adult
neurons , but , interestingly enough , it is known to be violated by
developing neurons in the very young . In any event, the accepted
classification specifies neurons as dopaminergicif they typically release
the transmitter dopamine , cholinergicif they typically release acetylcholine
, and
so on
for
the
set of neurotransmitters
so that
fluoresce
it will
, researchers
have
been
able
to show
areas
of
It was hoped that particular neurochemicals might be found exclusively in certain tracts and areas whose function was more or less well
defined , but these hopes have been disappointed . For example, it
would perhaps have been easier if the enkephalins were found solely
in tracts that carry pain information or mediate the pain response , but
this has not turned out to be the case. Though enkephalin is concen trated in certain spots , these spots are not what one would expect if it
were primarily a neurochemical for the nociceptive system. For example, it is found in the primary photoreceptors of the spiny lobster .
It is also noteworthy that endorphins are found in a wide range of
organisms, including leeches, spiders, lobsters, rats, and monkeys .
Additionally , it is evident now that there is no sharp division between
the
chemicals
found
in
the
brain
and
the
hormones
found
elsewhere in the body . Consequently , the once sharp distinction between secretion of peptides by glands in the endocrine system and
secretion of peptides at the synaptic terminal is fast losing its edge.
All the peripheral peptide hormones - for example, the gut peptide
cholecystokinin
the brain , and peptides originally believed unique to the brain - for
79
Figure 2.29
example, 3-endorphinare
feld 1980).
mitterserotoninhas
been shown in Aplysiato cause either excitation or one of two styles of inhibition on the postsynaptic neurons,
to function as transmitters.
To count as a transmit-
80
1
Synthesisof transmitters (T)
Release
(secretion
)
3
Binding
to receptor
(R)
3TL
: =J
Removal
or destructionof T
TE
'.",T,'"E
,,
Figure2.30
Four biochemicalstepsin synaptictransmission:(1) synthesisof the neurotransmitter
substance(T), (2) releaseof transmitterinto the synapticcleft, (3) binding of the transmitter to the postsynapticreceptor(R), and (4) removalor destructionof the transmitter
substance
. (Reprintedwith permissionof the publisherfrom Schwartz(1981). Ch. 10of
Principles
of NeuralScience
, ed. E. R. KandelandJ. H. Schwartz, pp. 106- 120. Copyright
@1981by ElsevierSciencePublishingCo., Inc.)
ter, a substance must satisfy four conditions : (1) it must be synthesized in the presynaptic neurons , (2) it must be released from the
presynaptic terminal , (3) it must be shown to cause EPSPsor IPSPs,
and (4) there must be mechanisms for its removal from the scene of
action (figure 2.30).
Demonstrating that a neurochemical passes muster on each of the
four counts is exceedingly difficult , and to date only eleven have
passedall four, though many others have been shown to passsome, and
answers are awaited for many tests. The list of canonical transmitters
includes acetylcholine (an excitatory substance released by motor
neurons at the neuromuscular junctions , and by other neurons in the
CNS as well ), dopamine (so far found to be inhibitory ), norepineph rine (also called noradrenalin ), serotonin , and GABA (-y-aminobutyric
acid, another inhibitory transmitter ) (tables 2.1, 2.2).
Not all inhibitory transmitters have precisely the same causal
profile , however . Norepinephrine appears to have a longer and
slower course of action than , for example, serotonin (Taylor and
Stone 1981). Some neurochemicals are thought to have a modulatory
role in regulating the effects of a transmitter on the receiving cells,
and there is evidence that some neurotransmitters act on voltagesensitive channels as well as at the standard receptor sites, thus affecting the response properties of neurons in a range of subtle ways
18
l
saJue
' G algEl
~ IEJ ! UOUEJ
bombesin
~sgns
la ~~! WSUEl
983 .
SUOlnaN Jo A1Oa41UlapOW
82 Some
Elementary
Neuroscience
(Nicoll 1982). The subtlety and complexity of interneuronal communi cation far outstrips the " excite-inhibit " repertoire , and we are only
beginning to understand what the assorted chemicals do (Taylor and
Stone 1981, Dismukes
1980) .
more
is known
about
the
details
of brain
business
and
about
, but
will
it be
solved
much
of
that
any clear idea of how they affected the CNS and were there -
Research now indicates that the phenothiazines block the dopamine receptor sites , and the evidence has come from the cellular as
well as the behavioral level . In vitro studies show that chlorpro -
83
84
SomeElementary
Neuroscience
85
pointed to a genetic effect: roughly 18 percent of the first group became schizophrenic , whereas the incidence of schizophrenia in the
second group and the control group was the same , about 10 percent .
1982.)
environmental
whereas
in an environment
free
of those
factors
a carrier
of the
gene
may be largely free of the disease. For example, being raised by parents who are schizophrenic appears to be one relevant environmental
condition
mined . It is worth bearing in mind that some genetically based diseases such as Sanfilippo 's syndrome have essentially one clinical
presentation but are producible by the presence of either of two distinct
genes
, and
for
all we
know
now
this
may
be the
case
for
schizo
1982 .)
of current
86
son' s disease, Huntington 's chorea, and myasthenia gravis) are also
instances where research in neurochemistry has been pursued with
considerable success (Spokes 1981). For wider and more thorough
discussions, the reader may consult works recommended in the reference section.
The discovery that it is possible to change the brain directly and
alleviate the catastrophic effects of brain disease has far-reaching social importance . Brain disorders are widespread , and demographic
distribution being what it is, the number of older citizens steadily
increases- with the result that the number of patients with strokes,
Parkinson 's disease, Alzheimer 's disease, and so forth , ,will also be on
the upswing . The social and personal costs of brain disorders are
enormous , and the possibility that we might really understand and
manipulate some of the causally relevant conditions seems an in creasingly real possibility . In this context it is useful to mention that a
new technique for treating Parkinson' s disease is being developed, a
technique that presents an important moral issue, especially since it
may conceivably be applied more generally to other disorders .
The issue is this . As a result of research in neural implants , it was
found that foetal neurons extracted from a specified area of a rat' s
brain and injected into the same area of another rat's brain established themselves there as functioning elements in the nervous system. Thus, neurons extracted from the substantia nigra of a foetal rat
and injected into the brain of a rat with lesions in the substantia nigra
continued to live and began to produce dopamine , the neurotrans mitter in deficit since the destruction of the substantia nigra . The
dopamine deficiency suffered by the brain -lesioned rat was accordingly made good to some degree, and the motor deficits were dimin ished though not abolished . Because the brain is isolated from the
immune system by the " blood -brain barrier " (the neurons are not
directly nourished by the bloodstream but are " fed" by the interven ing neuroglial cells), brain grafts are less susceptible to rejection by
the immune system as a foreign invader . Although this immunolog ical privilege , as it is called, has been demonstrated in the rat, it is not
yet known whether primate brains enjoy the same property and will
accept foreign neurons (Barker and Billingham 1977).
Now under study is the possibility that humans with Parkinson's
disease might be treated by injecting foetal neurons into the brain .
This would be far better than treatment with L-dopa, because the
brain 's need for dopamine is continuous but the drug can only be
taken at discrete intervals , and because in the long term Parkinson's
patients on this drug tend to develop psychiatric symptoms , some of
the motor deficits reappear, and the drug seems to have declining
87
treating diabetes with foetal cells from the pancreas, and it is possible
that replacing neurons in the hippocampus may help patients with
Alzheimer ' s disease. Foetal neurons implanted into the hippocampus
even show a tendency to establish projections to the appropriate
regions of the host brain , and more remarkably , they appear to establish functional synapses, though this is still under investigation
(Bjorklund
The most effective implant cells are foetal cells, and if it should turn
out that cell transplants
tuses
rather
than
from
foetus
~ s , the
moral
issue
comes
into
focus. So far no human foetal cells have been used in neural implants ,
and the experiments with implants in the brain of humans with Parkinson ' s disease have all used cells from the patient ' s own adrenal
be used
' s?
to alleviate
the
ravages
of
cruel
. Even
then , however
, those
who
believe
that
the human
soul
neuroscientists
taking any research that might provoke an imbroglio , however misguided its motivation . Thus, there is the further sociological and
political issue concerning how to educate the public such that a rational , nonsuperstitious
judgment
' s and
Alzheimer
' s disease
Although these questions are intriguing , I shall take them no fur ther in this book . My intent here is merely to flag the questions, partly
for their intrinsic interest , and partly because the issue of the use of
human foetuses will likely come to legislation in the near future . It is
an issue of profound practical significance, and one on which cooperation between neuroscientists and philosophers will be essential. I do
not believe there are any a priori answers , and the more empirical
88
89
so on .
In normal adult animals (rats and horses are well -known examples )
some point in his early development testosterone will be manufactured by his emerging testes and released in his body . Some will
affect maturation of the gonads, and some will reach his CNS where ,
during a certain critical period in which particular neurons are maximally sensitive to testosterone , it will induce organizational changes
in the brain that " masculinize " it . The critical period varies from
species to species . In most animals studied it is confined to foetal
stages , and in humans it is believed to encompass the third and
of nature
are one
indicator
of this
sort
of dissociation
. Freemar
tins are cows (genetic females) with a male twin . They are invariably
infertile
Even
ultrastructural
behave
more
studies
of
like bulls
brains
than
revealed
like
what
cows .
differ
of the
female
' s brain
as a result
of the
foetal
environment
she
condi -
90
measured by frequency of lordosis (female presenting postureraised rump , concave back). It is important to note that for such
animals, priming with testosterone after the critical period fails to
elicit mounting . Females exposed to testosterone during the critical
period are masculinized ; their adult behavior is characteristically male
and their lordosis quotient is low and not much enhanced by priming
with estrogen. The androgenization of the developing brain appears
to be irreversible and enduring .
Taking such standard behavior patterns as mounting and intromis sion as an index of characteristic male reproductive behavior and
lordosis as an index of female reproductive behavior , it is evident that
genetic sex determines brain sex only via the intermediary auspices of
gonadal hormones . Such indexes are admittedly rather restricted and
other behavior profiles such as display patterns, offspring -care behavior , song, nest building , etc., can, where appropriate , be taken
into
account
Establishing behavioral indexes for characteristically male or characteristically female behavior is notoriously difficult , and in the
pioneering stages researchers have tended to adopt fairly conventional and crude descriptions . How to extend animal studies to the
human case, how to refine and reconfigure the indexes of what is
masculine
and
what
is feminine
, and
how
to extend
studies
to include
of
the
masculine
and
feminine
in
behavior
, it is clear
that
and
that
this
exposure
determines
whether
exposure
as an adult to gonadal hormones will be effective in producing characteristic reproductive behavior . There also seems to be a distinct
feminizing of the brain , so that a female brain is not one that merely
missed out on androgenization (Goy and McEwen 1977, Feder 1984).
What exactly does testosterone do in the brain , and what does
brain androgenization mean in terms of neuronal structure and organization ? To begin with , rat studies show selected areasin the brain
where testosterone concentrates, and these areas include the hypothalamus , the pre-optic area, the amygdala, the midbrain , and the
spinal cord . Within these regions are specialized neurons that contain
protein receptors fitted for estradiol . Testosterone molecules enter
such cells , convert to estradiol , and link up with the receptors . The
protein -hormone pair migrate to the cell nucleus and enter it , where
they interact with the genes to affect the program for protein synthesis (figures 2.31, 2.32).
pac
91
hf
Figure 2.31
Abstract representation of a " generalized vertebrate brain ," showing locations of estradiol - and testosterone-concentrating neurons common to all vertebrates so far
studied . The top drawing is a horizontal view , and the bottom a sagittal view . Black
dots represent groups of steroid-concentrating cells. Abbreviations : a, amygdala; cbllm,
cerebellum; ctx, cortex; ht, nuclei in hypothalamus ; oc, optic chiasm; oIl, olfactory bulb ;
pit , pituitary ; poa, pre-optic area; s, septum . (From Morrell and Pfaff 1981.)
Paradoxically, it is estradiol , a female hormone , to which testosterone converts and which articulates the androgenization of the
brain . Why then , does the estradiol routinely produced by developing females not itself androgenize the brain ? The answer is that
in the foetus the liver produces a protective protein - alphafetoprotein - that binds estrogen in the blood and prevents it from
entering the cerebrospinal fluid and so from reaching the brain . Testosterone is not so bound and accordingly is free to find its way
unencumbered to the brain . This answer suggests the following test:
administer estradiol directly to the brain of infant female rats to see
whether it produces the characteristic androgenization seen in males.
The outcome is that it does indeed androgenize the brain of the genetic female (McEwen 1976, McEwen et ale 1974).
94
Cb
Figure 2.34
Drawing of a sagittal section of the brain of an adult male canary. HV c and RA are the
two forebrain nuclei involved in song control . L, the auditory projection of the caudal
neostriatum (N ), sends fibers that end in a field apposed to ventral HVc , presumably
enabling the learning of a song repertoire by reference to auditory information . nXII is
the caudal half of the hypoglossal nucleus, formed by the motor neurons that innervate
the muscles of the trachea and syrinx . Cb is the cerebellum . (From Nottebohm
1981.)
between steroid-sensitive neurons and reproductive behavior is beginning to yield impressive results . For example, male mating vocalizations in a range of different bird species are mediated by specialized
nuclei whose neurons are steroid binding . Nottebohm has found in
canary brains a pair of forebrain nuclei5 that are large in males and
small in females, and whose size can be manipulated by manipulating
the availability of testosterone (figure 2.34) (Nottebohm 1981, DeVoogd and Nottebohm 1981). In canaries only the male is a songster,
and Nottebohm has shown that well -developed nuclei are essential in
the canary' s learning and producing his song. Castrated males have
diminished song -control nuclei and cease to sing ; females ovarec tomized and given testosterone grow song -control nuclei and learn to
sing , albeit not with the customary virtuosity of intact males . Similar
95
96
97
tion of what we value, and this , finally , is a moral matter . All things
considered, I believe we should welcome research on sexual differ entiation in neuronal morphology , though the swill that can be
expected from the popular press will doubtless smother some enthusiasm . Still , it is better to know than not to know .
Selected
Readings
Akil , Huda , S. J. Watson, E. Young , M . E. Lewis, H . Khachaturian , and J. M . Walker
(1984). Endogenous opioids : Biology and function . Annual Reviewof Neuroscience
7:223- 255.
Bjorklund , Anders , and Ulf Stenevi (1984). Intracerebral neural implants : Neuronal
replacement and reconstruction of damaged circuitries . Annual Revie'lu of Neuroscience7:279- 308.
Bullock, T. H ., R. Orkand , and A . Grinnell (1977). Introduction to nervoussystems. San
Francisco: W . H . Freeman.
Cooper, J. R., F. E. Bloom, and R. H . Roth (1982). Thebiochemicalbasisof neuropharmacol
ogy. 2nd ed. New York : Oxford University Press.
DeVoogd , T. J., and F. Nottebohm (1981). Sex differences in dendritic morphology of a
song control nucleus in the canary: A quantitative Golgi study . Journalof Comparative Neurology 196:309- 316.
Feder, H . H . (1984). Hormones and sexual behavior . Annual Review of Psychology
35:165- 200.
Hawkins , Robert D ., and Eric R. Kandel (1984). Steps toward a cell-biological alphabet
for elementary forms of learning . In Neurobiologyof learning and memory, ed. G.
Lynch , J. L. McGaugh , and N . M . Weinberger , 385- 404. New York : Guilford .
Henn , Fritz A ., and Henry A . Nasrallah , eds. (1982). Schizophreniaas a brain disease
.
New York : Oxford University Press.
Iverson, Leslie L . (1979). The chemistry of the brain . ScientzftcAmerican241/3:134- 149.
Iverson, Leslie L ., and Martin M . Rosser (1984). Human learning and memory dysfunc tion : Neurochemical changes in senile dementia . In Neurobiologyof learning and
memory, ed. G. Lynch , J. L . McGaugh , and N . M . Weinberger, 363- 367. New York :
Guilford .
Kandel, Eric, and James H . Schwartz, eds. (1981). Principlesof neuralscience
. New York ,
Amsterdam , Oxford : Elsevier/North -Holland .
Shepherd, Gordon M . (1983). Neurobiology. New York : Oxford University Press.
Chapter 3
Functional N euroana tOInY
Understanding
the mind may not be as intricate
our intellect feared .
Rodolfo Llinas , 1986
3 .1
hoped or
In trod uction
If we are to understand
not
as our vanity
only
the nature
populations
activity
how
of neurons
permits
the
the brain
of the basic
are configured
organism
works , we must
to make
its way
their
in
understand
, but
the
also how
orchestrated
world
. An
impression
common outside of neuroscience
is that in their organiza tion , nervous systems resemble nothing
so much as an endless , un tended bramblescape - foreboding , hopelessly
tangled , and defiantly
intractable . Accordingly
, on the bramblescape
theory of brain design ,
figuring
out how the brain \l\Torks is a faintly ludicrous business , since
there are too many
The bramblescape
systems are indeed
structural
data will
and anatomy
is one of mutual
is brought
within
scientific
ken ,
emerge , which
in turn
spawns
new
func -
100 SomeElementary
Neuroscience
tional exploration , and so on . Moreover , the distinction between
structure and function is not , at bottom , a fundamental division in
kinds . Roughly , a concept is functional (physiological ) if it specifies
the job description ; it is structural (anatomical) if it specifies what
units in the machine perform the job . The distinction is a relative one,
however , inasmuch as concepts can be employed both at one level of
description to specify a function executed by structural units and at a
higher level of organization to give a structural specification for the
higher -level functional concepts. And at a lower level of analysis, the
hitherto structural units are characterized functionally relative to
the units that execute their lower -level function . As Christensen et al.
(1980) put it , " . . . one neuroscientist ' s mechanism is another' s phenomenology " (p . 342). This relativity certainly does not mean that the
distinction between structure and function is useless, but only that it
cannot be expected to shoulder too much metaphysical weight . (For
further discussion, see chapter 10.)
I shall touch on five dimensions of neuronal organization in this
chapter: pathways , laminae, topographical maps, columns , and nervous system development . In each case I shall provide only a few
examples to illustrate the organizational feature; for a more complete
introduction to the organization of nervous systems, see the selected
readings at the end of the chapter . There are other dimensions of
order , such as lateral inhibition and center-surround organization ,
which I shall not discuss here. To orient the reader, I will first in troduce the principal anatomical divisions in the human brain , rely ing on the figures to provide supplementary detail .
3.2 PrincipalAnatomicalDivisions
Based largely on features grossly observable in dissection, the tradi tional anatomical division of the brain distinguishes five main structures (figure 3.1): (1) the telencephalon, which consists of the cerebral
hemispheres, (2) the diencephalon , which consists of the thalamus
and the hypothalamus , (3) the mesencephalon or midbrain , (4) the
metencephalon, which is the cerebellum and the pons, and (5) the
medulla oblongata, sometinles called the " myelencephalon ." These
classifications are still used, though as progress is made in under standing how nervous systems work , how they evolved , and how
they develop , different classifications, functionally more revealing,
can be expected to emerge.
In humans , the cerebral hemispheres are very large (see figure 2.1)
and markedly wrinkled , as the expanding cortical sheet has folded to
stay within the confines of a skull small enough for the human female
Functional
Neuroanatomy101
Cortex
Cor
u~
,
,
ctau
~ trum
Plln \ - -
Ml ' dull ..
fourth
ventricle
B
S'-'ctiona ..
Majorsubdivisions
of thebrain
1. Telenc
'-'phalon
2. Diencephalon
3. Mesencephalon
4. Metencephalon
(ponsandl'crebellum
)
S. Myclenl
.:ephalon(mcdulloa
nblunJ
;tt;t)
obillng .. t..
Oor ~ 1 funil
' ulu ..
Dor ~ 1 horn
iltera
nil ' u lu ~
lilt
.:r .. )
horn
~
b
Vl ' ntril ) horn
S".".tionb b
Vent
rill funil
' ulu ~
Figure3.1
The centralnervoussystemconsistsof (A) the brain and spinal cord. Gray and white
matter are distributed as shown in crosssectionsthrough (B) the brain and (C) the
spinal cord. (From Heimer (1983). TheHumanBrainandSpinalCord. Copyright @1983
by Springer-Verlag.)
102 SomeElementary
Neuroscience
to deliver at birth . A fissure is called a sulcus, and the ridge between
sulci is a gyrus. The surface of the hemispheres is the cerebralcortex,
about 2 millimeters deep (figure 3.2), below which are white -matter
tracts (bundles of nerve fibers), the most renowned of which is perhaps the corpus callosum, which connects the two hemispheres.
What differentiates white matter from gray matter is this : the gray
matter consists mainly of cell bodies and dendrites and looks gray
when fixed in formaldehyde , whereas the white matter consists
mainly of axons covered with myelin , which has a whitish
appearance.
Major sulci and gyri in the cerebral hemispheres constitute a basis
for further anatomical divisions into lobes(figure 3.3). A more fine grained division was proposed by Brodmann (1909), who mapped the
brain into fifty distinct regions on the basis of cytoarchitectural criteria
(figure 3.4). (See section 3.4 for the explanation of his method .) The
Brodmann numbers are still commonly used in identifying cortical
regions, though in those regions such as 17 and 18 where physiolo gists have found functionally important differences within a specific
Brodmann area, new numbering systems have become necessary.
(See section 3.5.)
Subcortical gray structures important to note for later purposes
include the hippocampus and the amygdala, which are located within
the temporal lobe and are phylogenetically ancient (figure 3.5). These
structures, especially the hippocampus , are currently under intense
investigation for their role in memory . (See chapters 4 and 5.) There
are other large gray subcortical structures, such as the basal ganglia,
which figure importantly in motor control .
The thalamus , considered part of the diencephalon , is also a subcortical gray structure . Shaped like an egg and bearing complementary right and left halves, the thalamus is sometimes regarded as a
kind of brain within the brain . It is a major integrating center, and all
sensory tracts except the olfactory tract converge here before sensory
signals are processed and passed on to the relevant cortical area. The
lateral geniculatenuclei (LGN) are prominent thalamic structures, one
nucleus situated on each side of the thalamus, toward the rear. (By
" nucleus" is typically meant a well -differentiated clump of cell bodies
and dendrites ; in effect, the internal clumpy analogue of the outer
rind -like cortex.) The LGN are recipients of massive projections from
the retina and in turn project massively to specific areas of the cerebral cortex.
The midbrain (mesencephalon) contains a variety of structures that
function as integration and relay centers for incoming and outgoing
signals, and it is the connecting point for several cranial nerves con-
106 SomeElementary
Neuroscience
Fornix
Hippocampus
Amygdaloid
Figure
body
. 5
Schematic
diagram
pocampus
were
showing
the
transparent
1983
by
Springer
Deep
up
plexus
hydrocephalus
brain
positions
and
the
1983
CSF
structures
and
trickle
to
of
an
of
bathe
The
as
Human
YNithin
they
the
would
Brain
the
are
figures
and
CSF
cerebral
the
brain
be
Spinal
of
seen
the
if
Cord
ventricles
and
The
hip
the
brain
Copyright
near
sheath
prenatal
the
cooked
choroid
openings
in
in
filled
is
the
protective
CSF
cavities
CSF
called
during
of
results
ventricle
the
aqueduct
and
from
within
accumulation
abnormally
ventricles
emerges
brain
the
in
the
excessive
them
deep
fornix
structures
fluid
Blockage
expands
Heimer
specialized
. )
these
cerebellum
causes
( From
Verlag
cerebrospinal
by
relative
body
within
with
the
amygdaloid
of
the
dura
development
the
ventricles
condition
known
which
as
Functional Neuroanatomy
Interventricular
107
foramen
Lateral
ventricle
Third
ventricle
Choroid plexus
Cerebral aqueduct
Fourth
ventricle
II
,
II
Median
aperture
::
I
I
I
Figure 3.6
The brain 's ventricles . Frontal view of the brain , as if it were transparent , showing the
cavernous areas that are filled with cerebrospinal fluid (CSF). Three tiny apertures, a
median and two lateral apertures, permit the escape of CSF into the narrow space
between the brain and its membrane overcoats. (From Angevine and Cotman (1981).
Principlesof Neuroanatomy. New York : Oxford University Press.)
for the movement of your head . What is the circuitry that sustains the
tracking ?
To answer such questions , neuroanatomists have charted the path ways of neurons from the periphery , in the spinal cord , and in the
brain itself . Early research relied heavily on two techniques : (1) the
injection of special dyes that were taken up by neurons , and (2)
damaging a neuron , giving it time to degenerate down its length , and
then tracing the line of degeneration . Sometimes neurons postsyna ptic to the damaged neuron also show atrophying responses to the
damage , and this permits tracing the pathway further .
Recently it was found that radioactively labeled amino acids are
taken up by neurons , incorporated into proteins , and then shipped
108 Some
Elementary
Neuroscience
by axonal transport down the length of the axon. Sometimes the
labeled material
and is further
taken up by the postsynaptic neuron and shipped to its axon terminus . An enzyme , horseradish peroxidase , has the singular advan tage of moving in the other direction . Placed at the axon terminal , it
with
the anatomical
methods
(Heimer
and Robards
area
of the
cortex
to a well -
object smoothly even while the head is moving in any of its possible
parameters . The crucial part of the circuit for this reflex originates in
the vestibular
apparatus
in the inner
ear , where
vestibular
nuclei
in
the
brain
stem
and
receptor
neurons
then
to
the
oculomotor
nerves that direct the extraocular muscles to move the eyeball (figure
3.9). Therefore, there are only three synapses between the receptor
neurons
in the vestibular
apparatus
muscles
control -
110
A. SPINOTHALAMIC
PATHWAY
B. LEMNISCAL PATHWAY
BRAIN
STEM
LUMBAR
CORD
Figure3.8
Ascendingpathwaysof the somatosensory
system. (From Shepherd(1983). Neurobiol
ogy. New York: Oxford University Press.)
from the collicular route are the basis for the perceptual judgments of
blind sight patients (see chapter 5). The optic nerve carrying fibers
from each eyeball divides at the optic chiasm, with half the fibers
crossing over and half proceeding on the same side. As a result, all
the fibers representing the left visual field (originating on the right
hemiretina ) in each eye go to the right hemisphere, and all the fibers
representing the right visual field (originating on the left hemiretina )
go to the left hemisphere (figure 3.10). The first synaptic relay is in the
LGN , where fibers layer themselves according to the retina of origin .
In primates the fibers from the right eye occupy laminae 5, 3, and 2,
and fibers from the left eye occupy laminae 6, 4, and 1. (Laminae are
discussed in the next section.) After leaving the LGN , fibers proceed
in orderly fashion to specified layers in specified areas of the cortex. A
large number go to area 17, otherwise known as the primary visual
Functional
Neuroanatomy
111
~..-:;: :~: ~~
, ....-...-- - -~--~
-~.""""...----~-~
Lateral
Medial
Lateral
rectus
recti
rectus
I
-
C+:; I
\,~
I
.I
....,
' "
,-
"
"
"
;'
Granial I
"
' I '
"
nerve
nuclei ~ N. VI ' I :\ N.III '~II :' N.III ; ,\ N.VI I~
,
'"
"
"
I
,
Right
medial ,,i'
and
superior
vestibularnuclei"
-..
I
I
'
"
+ ': II:
"
I I\
I
....
'
\, and
Leftsuperior
medial
" vestibularnuclei
-
Scarpa's ganglion
I
I
I
Direction
of fluid
I
I
I
I
motion
Rotation
in the ampulla
I
I
Left horizonta '
semicircu
semicircular
jar duct
duct
Figure3.9
Schematiccircuit diagram for the initial phaseof the vestibula-ocular reflex arc. Increase( + ) or decrease( - ) in rate of firing along a particular pathway is indicated.
(Reprintedby permissionof the publisherfrom J. P. Kelly (1981b
). Ch. 35of Principles
of
NeuralScience
, ed. E. R. Kandeland J. H. Schwartz, pp. 406- 418. Copyright @1981by
ElsevierSciencePublishingCo., Inc.)
112
Figure 3.10
Projection
of theretinasuponthe superiorcolliculus
and thelateralgeniculatenucleus.
(Modified from J. P. Kelly 1981a.)
Functional
Neuroanatomy
113
114
2I -
Layers
of
striate
cortex
5
6
-
I
.
:1 y
X,
1 .,...-
LGN
r / ' ';
Magnocellular
V :'"
/
'
;"
x:j
I
y.,# ~
/ "
~
~
~.-";:'.,"
- -
To SUperiOr
.
coiliculus
o ptlC. tract
Optic nerve
LC/
~::::~ ~ ~
Figure 3.11
Schematic representation of the parallel organization of the pathways from retina to
cortex in the monkey . Three of the major morphological classesof retinal ganglion cells
are represented; the A and B cells are assumed to be the counterparts of the physiologi cally distinguished Y- and X-like cells, and the C cells to be the counterparts of one of
the subgroups of W-like ganglion cells. The X and Y cells project to the lateral geniculate nucleus (LGN ), the W- and Y-like cells to the superior colliculus (not shown ).
Within the LGN , X cells terminate in the upper (parvocellular ) laminae, Y cells in the
lower (magnocellular ) laminae . Their activities are separately relayed to striate cortex
(area 17), where they terminate in different laminae and activate different groups of
cortical cells. (From Stone and Dreher 1982.)
Functional
Neuroanatomy
115
Surprisingly , experiments by Sprague, Berlucchi, and their colleagues have consistently failed to bear out the prediction . Lesions to
areas 17- 18 result in diminution of visual acuity , but they do not
destroy the capacity to discriminate patterns or even to learn new
pattern discriminations (Sprague et ale 1977, Berlucchi and Sprague
1981). On the other hand , destruction of extrastriate areas does im pair pattern discrimination . These experiments clearly imply that extrastriate cortical areas make some independent contribution to
pattern recognition , and together with the neuroanatomical and
neurophysiological discoveries, they argue for a parallel-processing
model for vision . What functional principles might be implemented by
such a parallel architecture remain obscure, but some possibilities are
considered in chapter 10.
Also problematic for the classical hierarchical conception are recent
discoveries showing that the cortex has not merely one or two areas
whose cells are arranged in a topographic map representing the retina, but that it abounds in topographic maps. Topographic mapping
is the dimension of orderliness featured in the next subsection, and
the significance of multiple maps will be raised there.
It should finally be noted that another classification for pathways
uses neurochemical criteria , identifying a pathway according to the
characteristic neurochemical found in the axons of the group . For
example, the locus coeruleus is a small nucleus in the brain stem, the
axons projecting from which release norepinephrine (figure 3.12).
Axons from this structure are interesting because they project
throughout the brain , including allover the cerebral cortex, the spinal
~ord , and the cerebellum . This unusually wide spread of projections
has led to the functional conjecture that the locus coeruleus system
serves a regulatory role, perhaps in regulating attentional and arousal
states in the brain as a whole . Axons bearing dopamine , axons bearing serotonin , and so forth , are slowly being identified , and the
neuron groups in question are consequently sometimes referred to
as " the dopamine system" " the serotonin system," and so forth .
(See also figure 2.29.) The extent to which the neurochemical classification of pathways converges with other classifications is still
undetermined .
Though I have provided merely a glimpse in this section into what
is known about tracts and pathways , I trust it will nonetheless be
enough to fortify my claim that understanding structure is essential
for figuring out theories of function . (See selected readings at the
chapter end .)
116
Figure 3.12
Locus coeruleus norepinephrine system. (From Angevine and Cotman (1981). Principles of Neuroanatomy. New York: Oxford University Press.)
Functional Neuroanatomy
117
Functional Neuroanatomy
119
Layer I
-
II
-
Short corticocortical
-
IV
-
corticostriatal
corticorubral
corticothalamic (intralaminar)
corticobulbar
corticospinal and corticotectal
-I
Figure 3.14
Laminar distribution of somata of cortical efferent (pyramidal ) cells in monkeys . Verti callines to right indicate zones of termination in sensory cortex of thalamic (solid lines)
and corticocortical (interrupted line ) afferents. In areas outside the primary sensory
areas, thalamic fibers do not terminate in layer IV . (From Jones (1981). In TheOrganization of the CerebralCortex, ed. F. O. Schmitt et al. Cambridge, Mass.: MIT Press.)
3.5
liijiiiiiiiiij
lindrical aggregations of cells form rich interconnections . This columnar organization superimposed on the laminar organization will be
discussed below .
Such systematicity strongly indicates that the laminar structure is
the brain 's solution to a fundamental functional problem , or that
some sort of economy in information processing is achieved by such
precise stratification (Van Essen and MaunseII1983). Exactly what the
problem -solution pair is, however , has not been obvious . In this discussion the question of theory may be usefully delayed until further
dimensions of order have been discussed, since the various dimen sions of order may well be a packagesolution to certain functional
problems, rather than independent solutions to independent problems. (See also chapter 10.)
One of the most promising and puzzling discoveries about the or ganization of nervous systems is that many structures abide by a
principle of topographic mapping , whereby neighborhood relations
of cells at one periphery are preserved in the arrangement of cells at
other locations in the projection system . If we think of the neurons at
sensory
periphery
versions
of
regions
sheet
Whatever
successfully
since
more
all
In
the
order
ciple
is
maps
to
it
in
the
area
oughly
there
known
this
As
seen
of
other
up
the
cat
and
order
and
the
is
.
that
also
thor
of
Much
to
since
human
.)
in
the
Never
is
1959
is
clini
the
Roberts
less
mapping
and
monkey
cortex
applicable
Penfield
stimulate
experiments
topographic
in
most
human
ba
cells
to
The
species
the
the
species
relevant
from
strategy
crucially
systems
various
in
for
and
determining
record
between
indicate
for
sensory
the
prin
topographic
facts
downstream
organization
In
are
other
more
spatial
the
area
words
,
,
the
question
same
of
the
primates
( Allman
very
brain
is
map
the
cell
motor
cell
cerned
in
in
.
is
This
peripheral
the
as
the
retina
six
response
in
the
through
This
is
and
in
of
in
the
cell
is
cell
are
LGN
cortical
LGN
the
arrangement
in
ret
preserved
field
maps
in
the
fields
are
maximal
are
cells
of
receptive
executed
also
a
When
so
the
LGN
the
receptive
( The
a
field
sustain
retinotopic
move
that
in
touches
believed
visual
to
areas
of
) .
colliculus
level
on
.
functions
layers
adjacent
perpendicular
retinal
1977
superior
for
each
have
of
found
representation
evokes
all
stereoscopic
principal
eyes
that
that
been
topographic
neurons
,
so
terminating
also
relations
the
Additionally
part
subserve
the
spatial
field
has
cells
of
register
It
adjacent
the
cells
.
in
visual
. )
visuotopic
motor
laminae
organization
in
The
ganglion
six
arranged
specifically
the
retinal
in
lamina
most
systems
of
For
and
further
exists
of
arranged
the
as
turning
functional
methods
differences
chapter
have
neatly
served
Maps
we
each
CNS
be
well
portrait
periphery
motor
are
data
especially
Retinotopic
of
must
the
fuller
the
dimension
principle
( See
For
rarely
available
general
deformed
is
neuroanatomical
effects
interesting
justification
theless
animals
about
cal
it
as
what
standard
at
record
are
research
about
The
stimulate
studied
and
then
number
cortex
new
provide
exploit
study
and
and
techniques
to
under
centrally
its
to
systems
is
large
cerebral
theorizing
electrophysiological
strategy
in
the
arrangement
sic
a
served
the
maps
necessary
nervous
involve
sheet
in
be
topographic
of
such
constrain
will
receptor
principle
areas
in
time
represented
functional
wide
abound
forming
are
the
structures
ina
as
that
and
given
to
map
the
vision
on
sort
to
of
motor
that
number
the
quickly
grasp
map
is
to
foveate
visual
is
top
to
the
. This
be
stimulated
receptive
on
reflex
of
The
appears
perpendicular
eyes
a
the
and
periphery
layer
location
map
relevant
bottommost
the
the
structure
the
location
foveate
retinotopic
enables
laminar
of
a
,
field
stimulated
targets
dis
suggests
Functional Neuroanatomy
that
the
motor
map
extraocular
Different
the
classes
striate
is
to
level
in
ple
the
been
highly
found
been
ordered
show
on
continues
these
are
representation
the
is
field
not
yet
Allman
with
how
1977
as
similar
vein
capacity
and
figure
in
II
visual
but
together
field
in
stereoscopic
and
15
greater
others
of
the
ret
distortion
for
Merzenich
topo
have
map
design
the
research
of
than
of
15
distinct
systematic
center
as
areas
field
wraparound
emerged
and
which
some
the
the
another
interpret
the
figure
discovered
in
visual
the
vision
Kaas
make
these
split
such
such
in
as
as
representation
or
short
II
or
the
matched
Also
to
adjoining
and
moreover
between
highly
arise
order
transformation
minimal
distortion
and
also
representations
massive
cortical
and
II
projections
For
1980
: 26
responsive
and
area
connections
areas
colliculus
cortical
permits
reciprocal
various
layer
representation
order
arrangement
paired
the
with
regimented
from
first
This
superior
one
are
the
the
MT
are
between
typically
lor
between
finally
borders
and
second
around
thalamus
stimuli
wrap
relevant
the
congruent
functionally
visual
allows
DL
interconnections
between
projections
the
of
have
In
The
nections
to
that
around
best
argues
may
observations
areas
of
and
significance
example
wrapped
be
of
organization
ways
For
representation
II
evolved
and
settled
may
ten
retinotopic
functional
characteristic
parts
split
expanded
and
the
About
retina
complete
areas
central
is
means
the
distorted
for
some
no
addressing
are
maps
by
displaying
degrees
has
Califor
monomodal
has
the
the
multi
which
at
varying
basic
are
areas
of
to
more
there
hemisphere
and
contrast
topographic
appears
visuotopic
in
organization
inotopic
it
is
more
also
graphic
"
in
colleagues
each
in
matters
even
Important
that
representation
cortex
"
monkey
visual
of
at
his
markedly
topography
psychological
Research
cortex
be
out
distinct
complete
seven
"
the
respon
interpreted
owl
was
visually
more
the
infor
cortex
Other
turns
and
sixteen
to
and
remaining
Even
in
found
it
in
visual
striate
processing
In
Allman
the
laminae
how
cortex
be
and
of
specific
the
psychological
cortex
by
only
map
"
space
of
specific
to
the
to
less
as
Technology
been
have
project
Unpredictably
in
motor
conception
believed
studied
have
these
be
to
maps
of
LGN
retinotopic
to
hierarchy
Institute
areas
definable
the
perception
was
intensively
nia
classical
yield
which
retinotopic
the
the
referred
cortex
of
10
in
on
thought
sometimes
representation
chapter
cells
to
were
striate
See
and
have
areas
were
processed
thought
sive
of
cortex
mation
It
is
muscles
121
to
those
another
cortical
example
to
These
con
intracortical
the
opposite
Functional
Neuroanatomy123
'
.
-
.
.
.
.
- 60 ' ;
. .
.
.
- 30 ' .
. 10
0' - . . . . . . . . . . . . . .
1
O.
30' ;
:
.
60
;
; .
. .
.
B
Figure3.15 (continued)
(B) Schematicunfolding of the visual cortexof the left hemisphereof the owl monkey.
The unfolded visual cortexis approximatelya hemisphericalsurface, which is viewed
from abovein this diagram.
Abbreviations: AI, first auditory area; AL, anterolat"eral auditory area; CC, corpuscallosum; 01, dorsointermediatevisual area; OL, dorsolateralvisual area; OM, dorsomedial visual area; IT, inferotemporalvisual cortex (c, caudal; m, medial; r, rostral; p,
polar); M, medial visual area; MT, middle temporalvisual area; ON, optic nerve; OT,
optic tectum; PL, posterolateralauditory area; PP, posteriorparietalvisual cortex; R,
r()~tral auditorv
temporal
visual area; TP, temporoparietalvisual
~ area; ST, superior
cortex; VA, ventral anterior visual area; V-I, first visual area; V-II, secondvisual area;
VP, ventral posteriorvisual area; X, optic chiasm. (From Allman 1982. Reconstructing
the evolution of the brain in primates. In PrimateBrainEvolution
, ed. E. Armstrongand
D. Falk. New York: Plenum. Basedon Allman and Kaas1971.)
hemisphere arise from lower parts of layer III , and so forth . (Seeagain
figure 3.14.)
An additional organizational feature of the visual system is that as
cells frj)m the LGN project to specified layers of the striate cortex,
they do so in chunks alternately from one eye and then from the
other , producing " ocular dominance bands" of neurons in the cortex
(Hubel
, Wiesel, and LeVay
3.16). A similar sort of band,
- 1977)- (figure
- ing of terminating fibers has now been found in the prefrontal cortex,
where fibers terminating in the principal sulcus and originating in the
homologous area of the contralateral hemisphere form bands, with
Functional
Neuroanatomy
125
CochleotopicMaps
The cochlea, a structure of the ear, is a coiled tube containing specialized membranes, fluids , and organs. In virtue of the structures in the
middle ear, certain deformations in the air are transformed into wave
patterns in the cochlear fluid , which in turn are transduced into electrical signals in specialized receptor cells (figure 3.17). These signals
are transmitted up the auditory pathway to the cerebral cortex, and
the final consequence of these activities is that we hear sounds. (For a
basic introduction to the auditory system, see Lindsay and Norman
1977; for a more detailed account, see UttaI1973 .)
The cochlea is so constructed that high frequency tones stimulate
receptor cells at the base of the cochlea, and progressively lower tones
stimulate cells up the length of the cochlea, with the lowest-frequency
tones stimulating cells at the cochlear apex. The neighborhood relations in the frequency spectrum are preserved in the axonal layout in
the auditory nerve, so that a neuron with a characteristic frequency
of, say, 1500Hz will lie between neurons with characteristic frequencies of 1600 Hz and 1400 Hz .1 This preservation of neighborhood
relations is called a cochleotopic
or tonotopicmap, and the orderly mapping of neurons with sound frequencies is preserved at each synaptic
station on the way to the cortex: the cochlear nuclei, the superior
olivary nuclei , the inferior colliculi , the medial geniculate nuclei of the
thalamus , and finally , diverse areas of auditory cortex (figure 3.7).
Because each of these integrative nuclei contains divisions that congregate neurons containing distinctive source-specific information ,
126
Unwound
cochlea
b
Hair
cells
Basilar
membrane
Figure 3.17
(a) Position of the cochlea relative to the eardrum , showing the mechanical delivery of
acoustic energy to the base of the cochlea. This particular arrangement serves to match
impedances so that vibrations are efficiently transferred from the gaseous medium of
the atmosphere to the higher -density liquid medium within the cochlea. (b) The " unrolled " cochlea. Points along the internal membrane are mechanically resonant to
specific frequencies. Membrane motion at a resonant point stimulates the local hair
cells. Acoustic frequency is thus spatially coded in the nervous system. (Based on Yon
Bekesy 1960.)
Functional Neuroanatomy
127
this is not a simple serial processing system, but a system complemented with extensive parallel processing.
There are at least six distinct cochleotopic maps on the temporal
cortex of the macaque monkey , and four on that of the owl monkey ,
each with its own properties and its own source of thalamic neurons
(Merzenich and Brugge 1973) (figure 3.18). How many cochleotopic
maps are on the human temporal cortex is not yet known , but on the
basis of other primate studies it would be surprising if there were
fewer than six on each hemisphere . (In at least some animals there is
also a degree of auditory mapping in the superior colliculus .) Nor is it
yet known whether each representational region has a distinct func tion or exactly what such functions might be. There is also extensive
bilateral representation of sounds from the inferior colliculus on up ,
so that sounds detected by cells in the right ear, for example, are
represented in both right and left hemispheres. These intercalated
representations are believed to be the basis for the several binaural
" comparators" that integrate information from both ears to enable
localization of sound sources (Merzenich and Kaas 1980).
SomatotopicMaps
Sensory receptors in the hand enable us to feel light touch, pressure,
vibration , temperature , and pain , and also to know the position of the
hand and the joints . Distinct receptor neurons are sensitive to each of
these features, and neurons higher in the projection system preserve
the modality uniqueness .
Each peripheral neuron has a receptive field such that the neuron
will maximally respond when a particular part of the skin, for example, is stimulated with its preferred stimulus - say, light touch .
Size of receptive field varies, and some neurons , such as those in the
lemniscal system with receptive fields in the skin of the fingers , have
much smaller receptive fields than those for the skin of the back.
Greater perceptual acuity is achieved with a large group of neurons,
each of which has a small receptive field , than with a few neurons,
each of which has a large receptive field .
Neurons carrying information from the sensory periphery termi nate at selected areas of the thalamus, where they are arranged in a
somatotopic representation of the body . These thalamic areas in turn
project to selected areas of the cortex, all in very orderly fashion and
with substantial reciprocal projections from the cortex. In the cortex
neurons are again mapped in such a way as to preserve neighbor hood relations of body parts, though there are some interesting discontinuities . Accordingly , neurons responsive to stimuli on the index
finger are between neurons responsive to stimuli on the thumb and
128
Eo rI fer VIeW
liW
~ ca~ J8 Aud i tory Fields
Earlier
VIew
Cat AiJditory F~
Figure3.18
Traditional(A and C) and current(8 and D) viewsof the organizationof auditorycortex
in monkeysand carnivores. In all figures the representationsof the cochlearapex(A)
and base(B) are indicated in those fields where cochleotopicorganizationhas been
described.In the traditionalview, only primary (AI) and secondary(All ) auditoryfields
were described. Later, AI was redefinedand five additional auditory fields with distinct architectonicand electrophysiologicalcharacteristicswere described. Abbreviations: LAF, lateralauditory field; AAF, anteriorauditory field; PAF, posteriorauditory
field; VPAF, ventroposteriorfield; RAF, rostrolateralfield; NT, nontopographic; NTCMAF, caudomedialauditory field. FieldsA and B are shownon a view of the surface
of the superiortemporalplaneand dorsolateralsurfaceof the superiortemporalgyrus.
(From Merzenichand Kaas1980.)
Functional Neuroanatomy
129
those for the middle finger , but distant from those for the big toe or
the scalp. This organization yields somatotopic maps on the cortex of
the sensory surface, the muscles, and the joints . In general, the lemniscal system exhibits precise mappings of the body onto neurons of
the somatosensory cortex. The spinothalamic system projects to dif ferent areas, and here mapping is thought to be coarser and less
precise. Neurons in this system tend to have larger receptive fields,
and individual neurons may be receptive to an assortment of types of
stim uli .
Early research indicated one major somatosensory map, designated
as 51, and a smaller area, 511,.but recent research has revealed many
more complete somatotopic maps in the area behind the central sulcus (figure 3.19). Six have already been found in each hemisphere of
the owl and macaque monkeys , and the search is not yet complete .
Distinct somatotopically arranged areas of cortex appear to be specialized for distinct submodalities , such as joint position , cutaneous
stimulation , or " deep somatic" (muscle) parameters. The maps are
not simple duplications of one another, and there are important dif ferences between somatotopically mapped regions . The differences
range along such parameters as origin of terminating fibers, destiny
of projecting fibers, receptive field , and response properties of the
neurons . As with retinotopic and cochleotopic maps, it is suspected
that the human cortex will be at least as richly endowed with
somatotopic maps as are the already investigated species. And for
somatosensory processing, as for visual processing,
- the classical sequential model seems to be ceding ground to a parallel model (Kaas et
al. 1981).
One principle governing map distortion is based on the population
of neurons representing a particular body area. In the case of certain
maps some body areas such as the hands and lips are more represented than others, since those areas have a higher density of receptive neurons , which in turn have smaller receptive fields . It has been
found that the amount of cortex activated by a stimulus in a given
receptive field is roughly constant, whatever the peripheral area (Sur,
Merzenich , and Kaas 1980). This means that when a body part is
innervated by many neurons, each of which has a small receptive
field , that body part will appear magnified on the cortical map relative
to an area in which there are fewer neurons , each of which has a large
receptive field . Accordingly , a cortical map may show a much larger
area for the fingers than for the trunk .
Other distortions are not so readily explicable. For example, in the
owl monkey there are substantial differences between area 3b and
130 SomeElementary
Neuroscience
A
B ~Rostrocoudol
Bonds
"
Foot
Trunk
Hand
Face
3a 3b' 1 2
c
Multiple
Representations
lmm
Figure
3 , 19
Three
conceptions
of
monkeys
( A
of
,
The
and
traditional
fields
51
are
contains
left
and
retinotopic
.
The
in
does
hand
is
tion
the
vib
of
the
body
on
.
Portions
the
of
of
by
Closer
is
two
cutaneous
brain
in
the
on
the
relation
body
given
vibrissae
body
that
the
given
view
major
shown
to
parts
on
are
the
( From
four
indicate
the
view
are
mandibular
( 51
monkeys
that
Data
features
activated
,
is
( C
dorsolateral
cortex
vib
maps
numbered
vibrissae
bands
Somatotopic
of
are
of
owl
occupies
concept
in
50me
shown
right
cortex
in
representation
traditional
fields
are
the
hand
chin
single
the
1
on
maps
of
architectonic
and
mapping
much
than
the
that
somatosensory
representations
right
Kaas
et
aI
. )
area
of
3b
parietal
cutaneous
modification
detail
digits
:
been
all
auditory
Abbreviations
has
postcentral
two
representations
more
and
outlined
areas
in
the
has
multiple
in
lower
( 8
of
of
across
representations
1981
represented
region
organization
details
concept
architectonic
parts
the
some
the
split
map
between
the
hand
extensive
in
The
for
and
as
what
still
relations
ensembles
the
map
the
of
it
means
obscure
contain
the
in
area
glabrous
representation
and
is
topographic
neuronal
inasmuch
of
the
this
information
that
representation
area
reason
somatosensory
hunch
of
more
for
But
it
key
the
back
the
processing
is
to
3b
surface
hard
communica
of
to
the
stifle
Functional Neuroanatomy
Maps
in
Like
the
its
Motor
cohorts
adjacent
such
in
fingers
the
leg
neurons
has
the
superior
of
intensive
abstract
The
motor
of
In
as
plane
- like
into
Jones
tions
and
cortex
of
In
zone
other
lamina
thalamic
is
will
sory
cortex
umn
will
the
one
activity
with
cortical
organization
obvious
hitherto
is
and
more
suspected
, Schiller
is
in
the
)
to
and
of
to
Stryker
mm
in
the
the
1930s
dimension
that
in
of
cells
width
relationships
be
1972
horizontal
it
has
been
order
in
aggregate
.
The
and
, 1979
for
the
in
col
columns
intrinsic
, Mountcastle
these
bundles
fibers
originating
specific
some
zone
in
are
are
local
1979
cir
, Gold
the
largely
in
within
response
for
, or
other
the
fully
under
of
neurons
I this
about
the
say
circumscribed
properties
zones
and
areas
of
. As
termina
in
occupying
that
of
modality
terminating
a given
pressure
area
respective
homolo
, the
neurons
in
the
cortical
thalamus
means
,
originating
,
fibers
on
the
axons
thalamus
thalamic
simply
means
" punctate
imposed
. More
no
aggregation
the
in
information
neurons
related
to
degree
by
the
point
of
terminate
for
example
it ,
that
is
basis
hemisphere
carrying
and
columns
one
contralateral
to
in
these
termination
properties
are
have
in
a
1975
that
case
order
N6
appears
1 .0
- output
common
puts
the
displays
displays
, it
0 .5
evident
fibers
knee
and
) .
elsewhere
spatial
focal
less
areas
1972
de
also
input
role
it
( 1981
."
seen
3 . 20 ) ( Szentagothai
columns
area
in
Rosen
specifically
functional
gous
if
of
motor
, also
earlier
. As
on
surprising
strip
fashion
movement
re by
mentioned
laminar
research
Lorente
about
1981
, but
zones
cortical
motor
the
innervation
representation
eyeball
just
by
the
terms
- Rakic
focal
the
. As
pre
cortex
on
of
of
in
revealed
have
. More
( figure
The
stood
front
smaller
topographic
and
begun
bundles
in
cuitry
is
is
motor
density
in
it
be
the
to
ordering
. In
representation
higher
organized
not
were
we
that
defined
cord
topo
connected
) .
work
discovered
man
would
organized
are
Columns
cortex
vertical
( Asanuma
Vertical
is
4 ) is
cortex
topographic
greater
their
epithelia
maps
them
1972
plane
umn
it
spinal
controlling
sensory
and
Robinson
3 .6
contain
the
this
the
though
too
muscles
mapping
devoid
( area
of
6 , immediately
colliculus
to
and
to
layout
appears
,
of
owing
. Area
extraocular
cortex
regions
have
topographic
layer
motor
aspects
tongue
the
the
adjacent
muscles
motor
than
motor
of
the
and
gions
that
groups
served
of
System
sensory
graphically
131
on
the
set
the
of
left
somatosen
narrow
vertical
receptive
fields
col
.
Functional Neuroanatomy
133
134 SomeElementary
Neuroscience
Columnar and laminar
organization of cerebral cortex
II
III
IV
VI
Figure3.21
Columnar organization of the somatosensorycortex demonstratedphysiologically.
Eachcolumn is specificfor a submodality. The examplesshown hereare for joint 0),
pressure(P), and hair (H) stimulation. (Reprintedwith permissionof the publisher
from Kandel (1981c
). Ch. 17 of Principlesof NeuralScience
, ed. E. R. Kandel and J. H.
Schwartz, pp. 184- 198. Copyright @ 1981by ElsevierSciencePublishing Co., Inc.
Basedon Mountcastle1957.)
Functional Neuroanatomy
135
ACochlea
-~ n~ Cat
Auditory
Fields
(AAF
and
AI)
B. Biroxal Ba ~
within AI
I
I
I : . .)
--:
I
I
I
I
Figure 3.22
(A) Cochleotopic representation within two auditory fields (AAF and AI ) in the cat.
Numbers represent locations along the cochlear sensory epithelium measured from the
cochlear apex. (8) Highly schematized drawing of the internal organization of AI . This
field is marked by a series of alternating functional binaural bands, orthogonal to the
isofrequency contours of the fields . Within a given binaural band, all neurons are either
excited by the stimulation of both ears (EE bands), or excited by the stimulation of the
contralateral ear and inhibited by the stimulation of the ipsilateral ear (EI bands). The
lowest frequency part of AI (at the right ) has not been studied in sufficient detail to
determine its internal binaural organization . (From Merzenich and Kaas 1980.)
FunctionalNeuroanatomy 137
the macrocolumn enjoy their OWl) peculiar extrinsic connections to
modules
outside
the
macrocolumn
. As
Mountcastle
says :
sets of modules
of the
several
entities
are defined
as a distributed
system . (1979:37)
This connectivity arrangement implies that a microcolumn can belong both to a given macrocolumn and to a number of distributed
systems. This in turn would explain why local lesions typically do not
destroy a particular functional capacity, but rather tend to degrade it
to a degree roughly proportional to the size of the lesion and to the
general importance of the structure in subserving the capacity
(Mountcastle 1979) .
Much remains to be discovered concerning just which columnar
aggregates talk to which other columnar aggregates , and such infor -
standing what is going on in the brain at a number of levels of organization . It is, for example, a matter of considerable significance that
roughly 80 percent of an aggregate ' s connections - incoming and out -
going- are with other cortical neurons . Why cortical neurons have so
much
to say
to one
another
and
what
the
nature
of the
messages
is
comes to be understood
in greater detail .
to accommodate
the
data concerning pathways , laminae, maps, and columns . With in creasing degrees of accuracy we can think of the cortex as a stack of
sheets
, then
as a stack
of sheets with
methodical
, then
highly specificorigins and projections. Seeing so much architectural pur pose , one is virtually driven to conclude that the secret of the brain ' s
computational and representational capacities cannot be understood
Neural Development
The brain of an adult organism is remarkably orderly in its organiza tion , but perhaps what is most astonishing about such organization is
that no master intelligence guides its construction - it grows natu -
138 SomeElementary
Neuroscience
Figure 3.24
Diagrammatic illustration of some major events in the maturation of a Purkinje cell and
of the molecular layer. If the parallel fibers are traced through the columns from left to
right , it is seen that the fiber formed on day 7 has no synapses on that day, but that it
has synapses with basket cells on day 12 and an increasing number of synapses with
Purkinje cells from day 15 onward . Likewise , the parallel fiber formed on day 12 has no
synapses at that time, but has synapses with a stellate cellon day 15. The displacement
and growth of a climbing fiber is indicated on the left side of the Purkinje cell. (From
Altman 1972.)
rally , starting from a few cells induced from epithelial cells early in
the biography of the developing embryo . Studies in developmental
neurobiology have begun to reveal how strictly patterned is the
genesis of the nervous system, and how very puzzling are the princi ples that govern its development (figure 3.24).
Consider , for example, a motor neuron in the spinal cord of an
adult monkey . First, its causal ancestor in the embryo was a precursor
cell (neuroblast), from which it must have differentiated in a mitotic
division . This differentiation results in its having the distinctive prop erties of a motor neuron , not those of a mechanoreceptor or a Purkinje cell or some other type of neuron . Second, it must migrate from
the point of origin to the right place in the spinal cord. Then it must
grow its connections; the dendrites must sprout and make contact
with the right neurons to form the right circuits, and the axon must
find its way out of the spinal cord via the proper route (the ventral
horn , not the dorsal horn ). It must go to the muscles- more particu -
Functional
Neuroanatomv .I
139
if it is a ventromedial
axon
it must
innervate
the
axial
muscles
There is evidence that the various neuronal pools in the spinal cord
preferentially " seek" specified muscle fibers to innervate , a remarkable and almost uncanny route-finding accomplishment (Rakic 1975).
An equally extraordinary story can be told for all other neurons . For
example, a retinal ganglion cell must have ganglion cell shape and
response properties , migrate
from
to the most
where . Sometimes errors are made, but errant neurons are typically
eliminated as development proceeds. It is not known how errors are
recognized, or why the errant neuron fails to survive .
The electrophysiological properties of spiking neurons also
undergo a distinctive evolution . In their early stages the spiking of
these cells is based on calcium current . At later stages spiking may
have both a calcium and a sodium component , but eventually the
sodium component dominates and the calcium component declines .
This physiological sequence has not yet been fully explained , but it
appears to be an important feature of neurogenesis and may be espe-
more
(Cowan
1979 , Truman
robust
, a variation
on the
" them
that
1984 ) .
In addition to the culling of entire cells, there is in certain popula tions a massive retraction
140 SomeElementary
Neuroscience
cell, and in subsequent development all but one retract, presumably
through some sort of competition . Once matters have settled down ,
process proliferation in some cell types appears to resume.
Intervention in the normal course of neurogenesis has yielded valu able insights into the causal agents. A particularly illuminating technique pioneered by R. W . Sperry has been to transplant an extra
(supernumerary ) limb or to rotate an eye after cutting the optic nerve
and to study the nature of the subsequent innervation as well as the
behavior of the organism . Sperry found , for example, that if an extra
leg bud is grafted onto a chick embryo alongside the normal leg, it
will be innervated by motor neurons . Specifically, it will be inner vated by the motor neurons normally innervating the trunk (Sperry
1963). In studies on the frog in which the optic nerve had been sectioned and the eye rotated 180 degrees, he also found that retinal
ganglion axons will regenerate, but they follow the route of the origi nal axons back to the optic tectum . The behavioral consequence is
that the frog snaps 180 degrees aw~y from where the eye detects a fly .
This systematicity suggests some principled way in which neurons
determine where they are to terminate and with whom they are to
make contact. Sperry has hypothesized that a " chemical affinity "
exists between the growing neurons and the tissue of destination .
Some evidence supports this view , indicating that in forming connections , neurons are aided by protein " labels" acquired on their membranes in the differentiation process, in virtue of which they can
match with complementary neurons and be recognized (Cowan
1979). The development of a nervous system is plainly a matter of
great subtlety and complexity , and much more remains to be discovered about the chemical factors that affect neurogenesis and about
other mechanisms that may be involved . (For a discussion on the
developmental profile of the neocortex, see Rakic 1981.)
We have seen that the width of ocular dominance bands can be
influenced by suturing one eye shut and thereby depriving the relevant cortical area of activity . This implies that there is some stimulus dependent plasticity in the organization of the cortex and that within
the genetically delimited parameters there is scope for environmental
effects. How much of the organizational profile depends on the genetic program , as well as the manner of dependence, is under study ,
and one important discovery bearing on these questions has been
made by Martha Constantine -Paton and her colleagues (1981).
The background is this : in the frog the optic tectum is the dominant
visual processing structure , and such cortex as the frog brain has
plays no crucial role in vision . (In humans the superior colliculus is
the anatomical homologue to the frog' s optic tectum .) Unlike mam-
Functional
Neuroanatomy
141
second finding was that the fibers from the supernumerary eye and
the fibers from the normal eye compete for tectal space, with the result
that bands of tectal cells, sorted according to eye of origin , are formed
and interdigitate in a manner strikingly reminiscent of the ocular
dominance
bands
seen
in
mammalian
visual
cortex
. An
overall
ret -
positioned on the head, have visual fields that are not sufficiently
isomorphic . The supernumerary eye, though functional , actually im pedes performance on fly -snapping tasks, evidently because the retinotopic map in the tectum is not a faithful representation
visual
field
and
because
unusual arrangement
the
motor
connections
are not
of anyone
suited
to this
fibers
grow
to
the
tectum
rather
than
somewhere
else , and
a result
refinement
of this
of tectal cells .
additional
mechanism
, cells
are
maintained
as
RetIna
1Retina
2 ~o
@o ' P'V
Retina1 RetIna2
Functional Neuroanatomy
143
-,,.'-.
.':.',".
@@
Retina1 Refina2
OptiC
Tecta
8
Figure 3.26
Schematic explanation of banding when retinal ganglion cells from two eyes (retina 1
and retina 2) innervate one tectallobe . (A ) Complete overlap of the terminals from the
two eyes. Expected if only retinal ganglion cell-to-tectal cell affinities are involved in
generating the retinotectal map . (B) One of several ways the two retinal projections
could divide the tectal plate. Expected if ganglion -cell axons merely obtained some
aligning cue from the tectal plate and sorted so as to reestablish the nearest-neighbor
relationships of the retinal ganglion -cell bodies. (C) Interdigitating termination bands.
Expected if ganglion cell terminals from each retina are sorted so as to maximize contact
with particular regions of the tectum and maintain the nearest-neighbor relationships of
their retinal -cell bodies. (From Constantine -Paton (1981). In The Organization of the
CerebralCortex, ed. F. o . Schmitt et ale Cambridge , Mass.: MIT Press.)
Functional Neuroanatomy
145
Figure 3.28
Segmental ganglion seen from its ventral aspect, and action potentials of identified
cells. Cells labeled include the T, P. and N mechanosensory cells, which respond to
touch, pressure, and noxious mechanical stimulation of the skin. (Right)The intracellular recordings of T, P, and N action potentials were elicited by passing depolarizing
current through the microelectrode. The T cells fire repeatedly during a maintained
depolarization; N cells fire spontaneously and have a large undershoot. (From Nicholls
and Baylor 1968.)
essentially the same spike profile, and so forth (figure 3.28). Such
neurons, duplicated across individuals, are referred to as identified
neurons, and their presence, along with the relatively small number
of neurons in the total nervous system population, has made it possible to determine the entire physiological basis of behavior patterns
such as swimming in Hirudo (Stent and Kristan 1981) and habituation
and sensitization in Aplysia (Kandel 1979). In Hirudo about 20 to 30
percent of the roughly 200 neurons in each segmental hemiganglion
have
been
identified.
3.9
Conclusions
Anyone with theoretical proclivities will find in all this order not only
constraints that any theory of brain function must honor but also the
very fire to ignite the theoretical imagination. The neuroanatomical
information is sufficiently rich at this stage that the greening of theory
146 SomeElementary
Neuroscience
that articulates the principles of brain function seems inevitable . At a
minimum , those whose skepticism concerning brain research was
based on a bramblescape conception should be willing to see that the
detailed orderliness belies their presumption of impenetrable tangle .
The volume of research classed under the broad rubric of functional
neuroanatomy is enormous , and it is not possible to summarize even
the research devoted to a single aspect .3 Moreover , the order we
know about is in part a function of the techniques we have for making
it show up , and there may be further realms of order waiting for the
curtain to be pulled aside , had we but the appropriate techniques .
The order we know about is in part also a function of available
theoretical frameworks that characterize brain function and can guide
and motivate experimental investigations . A dimension of order may
be found only after a theory , robustly successful on a number of
counts , deduces that it ought to be there . With so much known about
the neuroanatomy and neurophysiology
of nervous systems , what is
now needed are theories of brain function that will begin to pull the
data together , to pose new questions , and to compete for epis temological space . This further matter of theory in neuroscience will
be explored in chapter 10.
Selected Readings
Barr, Murray L . (1974). Thehuman nervoussystem: An anatomicalvieu'point. 2nd ed. New
York : Harper and Row .
Cowan , W . Maxwell (1979). The development of the brain . Scientific American 241/
3:112- 133.
Heimer , Lennart (1983). The human brain and spinal cord: Functional neuroanatomyand
dissectionguide. New York : Springer-Verlag.
Heimer , Lennart , and Martine J. Robards, eds. (1981). Neuroanatomicaltract-tracing
methods. New York : Plenum .
Magistretti , Pierre J., John H . Morrison , and Floyd E. Bloom, eds. (1984). Nervous
system development and repair . Discussionsin Neurosciences
1/2 (special issue).
Merzenich , M . M ., and J. H . Kaas (1980). Principles
of organization
of sensoryperceptual systems in mammals . Progressin Psychobiology
and PhysiologicalPsychology 9:1- 42.
Schmitt , F. 0 ., F. G. Worden , G. Adelman , and S. G. Dennis , eds. (1981). Theorganization of the cerebralcortex. Cambridge, Mass.: MIT Press.
Spitzer, Nicholas C., ed. (1982). Neuronal development
. New York : Plenum .
Chapter
Higher
4
Functions
: Early
Work
I cannot conceive what even the very hiKhest nervous centres can possibly
be, except developments out of lower nervous centres, which no one doubts
to represent impressions and movements.
4 .1
Introduction
No claim about how the nervous system works can properly be assessed or understood without understanding the methods used to
obtain the data . This holds whether the claim is that dopamine is a
" science
. In earlier
sections
my
discussion
of methods
As things stand, there are many suppliant hypotheses, but few that
pass through the gates ; different
and
confusing results; there are deep problems concerning how to inter pret results ; the taxonomy of higher functions is itself ill defined and
theoretically impoverished ; in a few cases the conclusions researchers
148
where these are rife with guesses disguised as confirmed hypothe ses and flimsy suppositions masquerading as solid truths of
neuropsychology .
None of which should in the least discredit neuropsychology in its
attempts to get a grip on higher functions . For here the problems are
appallingly difficult , many methods are in their infancy , and any field
can be plundered by popularizers who hoist an hypothesis out of its
carefully surrounding guard of hedges, qualification , caveats, and
scare-quotes . Nor does it mean that neuropsychological
research can
safely be ignored , since masses of provocative data have been uncovered that are important pieces of the puzzle, and in any case,
neuropsychological questions will not be answered by sitting on our
hands and looking the other way .
In the main , research on higher functions in the brain has focused
both on questions concerning what parts of the brain carry out or
psychology on the hoof is much less straightforward . The central reason is this : even determining
joints .
Certainly we do not know a priori what the brain really does, nor just
what cognitive and subcognitive capacities it really has. We have to
find that out in the way we have to find out anything
, it will
be useful
to touch
on
them
now
Early Wark
149
memory ," and still others suggest a general distinction between " analytic " capacities and " holistic " or " synthetic " capacities . But what is
the psychological theory that provides the categorial framework and
guides the seach for neural substrates ?
The answer is that at this time psychological theory is in statu
nascendi. What is needed , but what does not yet exist , is a fleshed -out
and robust theory of what the fundamental cognitive capacities are ,
of the subcognitive capacities providing the lower -level basis for cognitive capacities , of the nature of processes that intervene between
input and output , and of the nature of the representations used at
various levels of organization . The starting point for theorizing was of
course folk psychology , just as the starting point for modern physics
was folk physics (see chapters 6 and 7) . Thus , according to our folk
psychological conception , we have a memory , we are conscious ,
some memories fade with time , rehearsing helps us remember , one
recollection sometimes triggers related recollections , and so forth .
Equipped just with the folk psychological notion of memory , we
might well expect that there is a single , unified kind of process or
mechanism in the brain that subserves remembering .
As psychologists and neuroscientists know well , folk psychology
does not take us very far ; moreover , the boundaries of its basic categories may have to be substantially redrawn . The investigation of
memory by psychologists has revealed a wide range of behavioral
data that has led to attempts to extend and redraw the basic folk
psychological conception . For example , it has been claimed that there
is a short -term memory , which stores information for short periods ,
and a long -term memory , which stores it more permanently . So far ,
however , these postulated capacities are understood essentially in
behavioral terms , not in terms of the well -defined internal processes ,
cognitive and subcognitive , that must underlie them .
Even for these ostensibly straightforward postulations , researchers
disagree on whether there are really two distinct capacities , because
the characteristics of the postulated memory stores vary as a function
of the materials to be recalled , the type of task set in the experiment ,
the strategies , and the subjects (Craik 1984). They also disagree on
whether there is a third element such as the rehearsal buffer (Atkin son and Shiffrin 1968) between the other two , how the various struc tures are related , and what the subcognitive processes are in virtue of
which people remember and forget . Moreover , remembering is not
self -evidently a single kind of process : free recall gives results different
from cued recall , which in turn gives results different from recogni tion tests (Craik and Lockhart 1972; Squire and Cohen 1984). In addi tion , it has been argued on the basis of careful animal studies that
150 SomeElementary
Neuroscience
working memory is dissociable from reference memory (alton , Becker, and Handelmann 1979) and that spatial-map memory is dissociable from nonspatial memory (O'Keefe and Nadel 1978).
Some data discovered by neuropsychologists are so remarkable,
and so contrary to customary assumptions, that they suggest that
some basic assumptions about memory may be in need of radical
revision . Consider , for example, patients who are so profoundly amnesic that they cannot remember the doctor they have seen day in
and day out , or what they had for breakfast, or anything of a close
relative's visit earlier in the day . Yet these patients can learn some
quite complex things , such as how to do mirror reading or how to
solve the Tower of Hanoi puzzle shown in figure 4.1 (though they do
not remember that they have encountered the puzzle before or that
they have learned to solve it ) (Milner , Corkin , and Teuber 1968,
Squire and Cohen 1984). Moreover , their memory for events that
occurred before the brain damage (retrograde memory ) is quite normal . Pertinent to the matter of a categorial framework for psychologi cal functions is the fact that there is as yet no principled description
specifying what general class of thing these amnesic patients can still
learn and what they cannot, and why they remember certain things
and not others . So far, no theoretically grounded description has
been winnowed out to specify the nature of the two capacities, if such
indeed there be.
At first , it might be thought that if distinct capacities really exist,
they might be distinguished as follows : one involves knowing how,
and hence concerns the acquisition of motor skills, and the other
involves knowing that, and hence is a matter of the acquisition of
cognitive information . That will not do, however , becauselearning to
solve the Tower of Hanoi puzzle is not merely acquiring a motor skill ,
such as learning to ride a unicycle (if that itself is a purely motor skill ).
It also requires something "that we might , in our innocence, call
in telligence .
Insofar as the learning in these cases has a cognitive dimension ,
one cannot just explain it on the basis of a cognitive -noncognitive
division . Indeed , these casesraise fundamental and disquieting questions concerning what it is for a process to be a cognitive process at
all, what processesunderlie skill acquisition , and what it is for behavior to be intelligent . (Also see chapter 10.) What makes the amnesic
studies so intriguing is that it is far from obvious what unifies the
assorted things that profound amnesics can still learn . From examining the class of things so far discovered that amnesics can remember,
it is not obvious how to predict what else they can remember and
Early Work
151
Figure 4.1
The Tower of Hanoi puzzle used to test cognitive skill learning in amnesic patients . To
solve the puzzle , subjects must move the five blocks shown in the top panel to the
rightmost peg. The rules are that only one block can be moved at a time, and a large
block can never be placed on a smaller block . The optimal solution requires 31 moves.
The middle panel shows an arrangement of blocks that leads directly to the optimal
solution , and the bottom panel shows an arrangement that does not lead directly to
solution becauseit is not on the optimal solution path . (From Squire and Cohen (1984).
In Neurobiologyof Learning and lvfemory, ed. G. Lynch , J. L . McGaugh, and N . M .
\JiJTeinberger
. New York : Guilford .)
what they will not be able to learn . Which entails that we do not yet
know what the capacities are (Warrington 1979).1
Learning and memory are at the de~d center of cognition , if anything is, and as their categories are revised and redrawn , the theoretical landscape of higher functions is undergoing tremenqous
transformations . The general category of learning has already fragmented into a variety of kinds of process, and indeed the term " learning" is now often replaced by the broader and less theoretically
burdened expression " plasticity ."
Among kinds of nervous system plasticity already believed to be
distinct phenomena are habituation , sensitization , classicalcondition -
152
Early Wark
153
154 SomeElementary
Neuroscience
are
usually
ogy
medical
"
is
serious
of
the
Ph
physiology
used
cludes
"
and
cover
neurobiologist
work
is
in
whose
Patently
is
the
neuroscience
in
sense
lists
Rather
for
with
just
it
although
brain
term
occa
sometimes
is
in
perhaps
informed
less
than
behavior
less
relationship
called
"
of
neurobiologists
others
"
who
are
And
still
are
others
who
of
the
are
try
to
use
cover
those
broadly
con
neuroscientific
more
others
to
cognitive
precise
widely
by
called
from
psychologists
very
call
people
latter
themselves
plain
called
from
some
matters
and
as
results
and
it
orderly
may
set
it
has
be
of
that
as
distinctions
eye
My
can
see
methods
literature
and
sometimes
at
in
every
is
that
and
selected
more
direction
and
with
that
levels
of
there
summary
might
the
it
mind
investigation
to
result
be
papers
shall
sometimes
sin
in
discussion
enable
greater
what
strat
seem
As
the
need
for
of
in
of
evolutionary
results
co
regretfully
providing
'
fruitfully
and
by
all
end
either
theory
vast
will
the
fulfill
unified
of
is
knowingly
hope
psychology
area
section
at
attention
importance
this
this
already
my
for
the
in
section
that
caught
prospects
indicates
literature
strategy
readings
texts
what
and
the
the
the
governing
selected
the
omission
my
The
neuroscience
particular
4.2
or
evolves
survey
how
inspect
it
and
is
themselves
with
general
choose
for
Moreover
norance
and
psychology
that
"
are
concerning
or
far
to
are
"
the
distinguish
neuropsychology
shall
suggests
as
applied
behavior
psychology
distinguish
to
concerned
keeping
be
chapter
by
in
Finally
egy
used
typically
and
there
to
like
constrained
in
brain
first
and
undoubtedly
but
divisions
emerge
Neuropsychol
Neuropsychologists
those
neurobiologist
is
no
and
cognitive
and
on
are
departments
research
ceived
will
II
neuropsychologists
"
is
behavior
there
"
"
term
"
university
from
psychology
are
the
but
neurobiologists
psychobiologists
who
and
indistinguishable
"
are
in
works
neurophysiologists
who
It
. "
ground
brain
typically
species
themselves
brain
Physiological
same
of
1949
human
found
"
been
nonhuman
whose
in
hospitals
having
neurology
the
studies
term
in
Hebb
often
animal
clinical
are
to
fashionable
the
departments
sionally
"
than
in
vintage
between
focus
usually
relation
wider
work
recent
by
of
has
and
more
publication
studies
"
doctors
term
the
circumspection
ig
of
reader
to
Early Work
155
a blockage of a blood vessel that cuts off the blood supply to a particu lar area of the brain or of a rupture of a blood vessel causing interrup tion of the blood supply downstream . If there is no blood supply ,
there is no oxygen, and if there is no oxygen, neuronal death swiftly
follows . Strokes are common in advanced age, and the nature of the
behavioral deficit followin ~ a stroke has been found to be linked to
the area of the brain damaged by oxygen deprivation . Study of stroke
victims has become an important source of information concerning
specialization of function , and it was also one of the earliest sources.
Nevertheless, behavioral deficits do not wear their causes on their
sleeves,
and for early physicians the study of deficits was confounded
.
In many ways .
For one thing , behavioral deficits can be caused not only by strokes
but also by tumors , traumas, infectious diseases, degenerating nervous diseases (e.g., Parkinson' s disease and multiple sclerosis), psychiatric diseases (e.g., schizophrenia ), and neurogenetic disorders
(e.g., Huntington 's chorea). Hence, it would have been exceedingly
difficult for early physicians to find much pattern in the medley of
symptoms their patients presented. Moreover , comparison of cases
was limited from one geographical region to the next, and testing
could be neither thorough nor standardized . Until rather recently ,
then , neurological diagnoses were largely intuitions in the dark . For
example, paralysis resulting from a stroke is not self-evidently distin guishable from paralysis resulting from a tumor , from excesseating of
a certain chick-pea (lathyrism ), from hysterical paralysis, or from a
demyelinating disease. Moreover , neurologically based disorders
such as strokes and tumors often involve personality changes, sometimes subtle, sometimes not so subtle . Even itemizing the roster of
possible causes, let alone isolating the particular cause of a particular
patient ' s malady, is knowledge hard won .
Disturbances in speech resulting from strokes are rather common ,
owing to the manner in which the cerebral blood vessels branch
(figures 4.2, 4.3). The middle cerebral artery feeds a wide area of the
frontal cortex, and interruption of blood flow in the left middle cerebral artery is often followed by loss of speech or aphasia. So far as is
known , the first person to make the connection between frontal cortex and speech production was Franz Joseph Gall (1758- 1828). His tory remembers Gall largely for his phrenology or " cranioscopy,"
as he called it , and his reputation as a crackpot has overshadowed
his deserved reputation as a fine anatomist and a pioneering
localizationist .
U sing better dissection techniques than others, Gall made postmor tem inspection of as many brains as he could get his hands on, where
156
Some Elementary
Neuroscience
"' "
.
,
"
--
-....
Anterior cerebralartery
Posterior
'Middle cerebralartery
communicatingartery
Posterior
- _~____r
Basilarartery
cerebralartery
Vetebral artery
External carotid artery
Subclavian artery
Arch
Figure 4 .2
The blood vessels
sclerosis
and
of Medicine
Sa unders
he
of the brain
obstruction
cerebral
areas show
the common
vessels . (From
, ed . P . B . Beeson , W . McDermott
, and
McDowell
sites of arterio
( 1979 ) . In Textbook
J . B . Wyngaarden
. Philadelphia
.)
had
documented
before
death
mass
must
itself
be
in
impaired
sulted
the
. He
ferentiated
thus
. The darkened
in the
of the aorta
patient
became
subserving
divided
' s capacities
convinced
a soul
into
with
distinct
behavior
, dispositions
that
the
distinct
" organs
. He
put
brain
the
was
, and
not
faculties
,"
damage
matter
of
traits
an
undif
, and
to
brain
that
which
it
re -
localization
:
. . . but
it
would
absolutely
so
different
intellectual
and
O ' Malley
be
also
homogenous
and
give
faculties
1968
quite
inconceivable
throughout
rise
so
:477 )
to
manifestation
various
that
could
and
a single
present
of
dissimilar
moral
organ
phenomena
qualities
. ( 1812
; in
and
Clarke
158 SomeElementary
Neuroscience
celJll
0" ,
Or',-.
~~
Q-ti
4"
-,
-;:'
"
.)'
0CJ
.-
.-:
. ,
.~~-~ :
..:;~:~:~.
.,~ "..:.:::
.~ .,~ .. r.-:.:
A ~ .;'.~ .-.
. ,, ~ '; ~ :i:.
. . .., .; - - ~ - - . .
, . . . ~ :"
Figure 4.4
Phrenological location of faculties according to the scheme of Spurzheim . (Reproduced
in Clarke and O'Malley 1968.)
The phrenology
came in with Gall ' s belief that superlative
develop ments of a capacity or trait required
superlative
development
of the
relevant brain center , which in turn would be reflected in the shape of
the skull as the enlarged neural mass grew against the cranial cap hence , the infamous
bumps
on the head . He thought
that sexual
appetite , for example , was a function
of the cerebellum , and hence a
consuming
appetite could be detected by feeling a telltale lump on the
lower rear of the head (figure 4 .4 ) .
, The bumps were both the glory and the undoing
of Gall , and he
and his followers , notoriously
Spurzheim , were often the deserving
objects of derision , in part because their overweening
zeal in fishing
for correlations
between character traits and cranial bumps overcame
any dispassionate
interest
in testing
their claims . Pierre Flourens
(1784 - 1867 ) was perhaps their most devastating
critic , and as a result
of his lesion studies on chickens and pigeons , he disputed
not only
the bumps theory but also the more general principles
of the localiza -
Early Work
159
Early Wark
161
162 SomeElementary
Neuroscience
time , Jackson well knew that surgical removal of a chunk of cortex
might yield no discernible effects.
Hughlings Jackson was profoundly influenced by Herbert Spencer's
evolutionary conception of nervous system function ,4 and he came to
view the brain as an integrated set of systems organized in a hierarchy, with sensorimotor representation featured at every level but
with increasing complexity and sophistication . Accordingly , his way
of reconciling the discrepancies between the " discrete organs" hypothesis and the " holistic " hypothesis was first to replace the exact
margins espoused by organ theory with imprecise boundaries and
then to argue for multiple representation in various places at various
levels of the hierarchies . This implied a division of labor in the nervous system , but a division made many times over , a division that
was fuzzy , overlapping , partially redundant , and increasingly specialized ; and moreover , a division of labor that had the potential for
remaIn
Early Work
163
information
about
frontal
lobe
function
came
from
animal
effects
of frontal
lobe
lesions
in animals
. There
was
of course
no
the surgeons
Walter
Freeman
and J. W .
effects of this
and
Thus, in addition to the problems already mentioned was the perplexing question of how to interpret a link between a lesion and a
behavioral
deficit , a question
164
the patient can no longer do Y, then A is the center for Y" ? For many
reasons, the answer must be no.
For one thing , the area A might be necessaryfor the function Y,
without being both necessary and sufficient as the description I' center" implies . Moreover , lesions in area A might result in interference
with other brain areas that are critical for Y, without A itself being
either necessary or sufficient for Y. It is known that acute lesions may
disturb functions elsewhere in the brain (Heilman and Valenstein
1979), and a behavioral deficit may be the result of such secondary,
nonspecific causes (also called " diaschisis" ). Finally , there might be
no " center" for Y at all . (For a general discussion of the problems in
interpreting lesion data, see Kolb and Whishaw 1980; Smith 1984.)
An illustration of this last situation can be found in a famous report
by Dejerine in 1892. He had a patient who awoke one morning and
discovered that he could no longer read, though he could speak and
understand . He could even write , but he failed utterly to read what
he had written . In addition , he had lost vision in his right visual field .
This syndrome is now referred to as alexia without agraphia, or pure
word blindness . If we may infer a center for every function that can
be selectively impaired , then it could be concluded that Dejerine' s
patient had a lesion in his " reading center." (Dejerine himself did not
draw this conclusion , but rather thought alexia to be a version of
aphasia.)
That conclusion is doubtful for several reasons. First, anthropolog ical considerations indicate that reading is a cultural achievement that
emerged long after humans had evolved . Second, since reading requires visual and linguistic capacities, and since postmortem examinations of patients with pure alexia typically show lesions both in the
visual cortex and in the corpus callosum, it may be more reasonable to
suggest (Geschwind 1965; Sperry and Gazzaniga 1967) that the deficit
in Dejerine's patient may be the result of a disconnection between the
relevant " linguistic " and " visual " areas of the brain . (For more on
alexia without agraphia, see Vignola 1983.)
Notice too that breezy use of the aforementioned simple inferential
scheme will yield a bizarre catalogue of centers- including , for example, a center for inhibiting religious fanaticism, since lesions in
certain areas of the temporal lobe sometimes result in a patient 's
acquiring a besotted religious zeal. Add to this centers for being able
to make gestures on command , for prevention of halting speech, for
inhibition of cursing and swearing, and so on, and the willy -nilly
nature of Gall's phrenological catalogue comes back to haunt us.
The interpretation of lesion results is further complicated because
function is frequently at least partially restored after a time, but the
Early \\ITark
165
166 SomeElementary
Neuroscience
beauties of the method is that experiments
can be undertaken
with -
. . . is endowed
with
electric
decided to go under the skull and investigate the brain more directly .
U sing dogs as their experimental animal , they surgically removed
segments of the skull . They then placed electrodes on the exposed
In a report published in 1870 they claimed they had obtained contractions of specific muscle groups from stimulation of specific areas
of the cortex , that the relevant area for a given muscle group was
small , that it was constant from dog to dog , and that contraction of
muscle
groups
correlated
with
which
areas
of the
cortical
con -
as motor
Because
other
cortex
cortical
areas
can
also
induce
movement
when
elec -
sensorimotor
representation ; wide
motor
cortex
, which
he called
a " motor
center
, " and
he took
his
EarlyWark
167
cancer of the skull that had eroded the skin and bone, leaving parts of
her brain exposed. Bartholow inserted electrodes into her brain and
was able to induce focal convulsions and sensations. For example, he
induced contractions of the upper and lower extremities on the right
side after stimulation of the left cerebrum. His account of this experiment makes chilling reading , as Mary is described as complaining of a
very strong and unpleasant feeling (Bartholow 1874) in her right arm
and leg. Bartholow was roundly condemned for using techniques on
a patient that were not motivated by clinical and therapeutic considerations, and consequently he languished as a medical outcast. Nevertheless, the work did show that a human brain could be electrically
stimulated to produce effects related to those found in animal
experimen ts.
Ferrier believed that his work with primates suggested that in humans too the motor cortex was crucial for voluntary movement and
that documentation on natural lesions and on epilepsy fit together
with his animal studies (Ferrier 1890). It was an inferential leap with
attendant uncertainties , but the evidence was decidedly intriguing .
Becauseresearchers were understandably reluctant to use humans as
subjects for electrical stimulation experiments unless there was a clear
clinical justification , the possibility of using this mapping technique
on humans to understand more of higher functions remained largely
in abeyance until Wilder Penfield saw in it an invaluable clinical tool ,
and one that might thereby have theoretical spin-off . Of this, more in
section 5.4.
Finer and more detailed examination of the motor and sensory
cortex in animals was conducted by C. S. Sherrington (1875- 1952)
and A . S. F. Grunbaum (1869- 1921),5 who not only improved upon
the mapping of the cortex but also delineated the difficulties and
limitations of electrical stimulation as a method , some of which remain with us still (figure 4.6). They noticed that if an area of the cortex
is stimulated at a particular time , this alters the response characteristic at that point and elsewhere in neighboring cortex, sometimes
facilitating responses, sometimes resulting in reversal (from extension to flexion of a joint ), sometimes producing a different movement
altogether . The borders of areas as determined by electrical stimula tion may vary as a function of what is stimulated when . As Griin baum and Sherrington ruefully point out in discussing the motor
field :
Thus if the anterior border is delimited by- stimulating- a series of
cortical points in succession from behind forward , the anterior
limit of the field is found farther anterior than if determined by
168 SomeElementary
Neuroscience
Early Wark
169
Chapter
Higher
Functions
Neuropsychology
and
Neurology
Apart from the technical and methodological difficulties , perhaps the most
serious obstacle in the way of real advance in the study of those aspects of
cortical activity that are commonly (if obscurely) referred to as " higher
functions , " is the absence, in most of neuroscience, of a sure theoretical
framework.
W . Maxwell Cowan 1981
5.1 Introduction
Results from lesion studies and from electrical stimulation studies
made it evident that the organization of the brain conformed neither
to the strict localization model nor to the holistic model, which took
degreeof activity as opposed to areaof activity as the important vari able. The strict localization hypotheses were undermined by a num ber of findings , both clinical and experimental . For example, in
experiments with monkeys Glees and Cole (1950) determined
through stimulation studies an area of the motor cortex specific to
movement of the thumb , then surgically removed the area and allowed the monkey time to recover from paralysis. During this period
the monkey recovered motor function of the thumb . When Glees and
Cole reexamined the cortical area, they found that cells bordering the
lesion, whose stimulation hitherto did not move the thumb on stimu lation , now moved the thumb . These too were lesioned out , and the
monkeys again recovered the capacity to move the thumb . These
studies, like Lashley's (1933, 1937), did not show there was no cortical
specialization for function , but only that strict localization of function
was doubtful .
On the other hand , quite specific effects resulting from brain lesions have been observed both in clinical studies and in the laboratory . For example, monkeys with hippocampal lesions show deficits
in spatial learning (Olton , Becker.1 and Handelmann 1979), and humans with bilateral lesions to the hippocampus , amygdala, and tem-
172 SomeElementary
Neuroscience
porallobe show profound anterograde amnesia (Scoville and Milner
1957) .
To illustrate the specificity of particular cortical areas, consider
color agnosia, studied intensively in the patient W .K . by Kinsbourne
and Warrington (1964). W .K . suffered a highly circumscribed lesion
to the left posterior cortex, in an area on the borderline between
visually responsive areas and areas subserving language. Tested once
his condition stabilized, he was found to display a largely isolated
deficit in color representation . He was unable either to identify correctly the color of objects or to pick out objects in response to a color
named. Nor could he correctly identify the color of the sky, trees, or a
stop light . The exceptions here, remarkably , were " white ," " gray,"
and " black." For example, he might say that a stop light is blue or
purple , and his responses were no better than chance. On the other
hand, he was unproblematically able to match and sort colors, such as
pieces of wool , to say whether two samples were the same or differ ent in color, and to say of a picture with a green sky that the sky was
the wrong color, though his performance in naming the color of sky
was at chance. His nonverbal representation for colors seemed, therefore, intact , whereas his verbal representation for colors appeared to
be disconnected from the nonverbal . He remained largely unaware of
the deficit , presumably because, as Kinsbourne and Warrington observed, " like a sensory aphasic who cannot detect the errors in his
own speech, the patient has no means of discovering his mistakes,
and he therefore continues unhesitatingly in his wrong use of the
names of the colors" (p . 299).
As another example, consider Mrs . B., a patient also studied by
Kinsbourne and Warrington (1963a). As demonstrated at autopsy,
she had a small, highly circumscribed lesion in the border area between the inferior temporal lobe and the occipital lobe. Her clinical
pattern showed a highly selective deficit : she had difficulty in perceiving the whole of what was shown , though she could , one by one,
perceive the parts. For example, if shown for a brief period a set of
geometrical forms , she could identify one, usually the left , well
enough, but her simultaneous perception of the whole scene was
very limited . Given enough time , she could identify other forms in
the picture . Shown a word , she could identify individual letters, but
she could not read the whole word . By reading each letter aloud ,
however , she could then reconstruct the word and by that laborious
means identify it . The impairment could not be attributed to a defect
in vision or to general intellectual deterioration , but appeared to be
specific to simultaneous form perception .
Like Dejerine 's (1892) patient (chapter 4), Mrs . B. was able to write
dictation
, but
she could
neither
copy
written
sentences
and Warrington
nor
173
read
suggest that if
other patients reportedly suffering from pure alexia were tested for
nonverbal form perception as well as for reading, they might be
found to be like Mrs . B. in having this more general deficit involving
simultaneous
least, it was very clear that the task-specificityof distinct areas was not
to be understood
on
the
model
of the
task -dedication
of distinct
seg -
after
certain
kinds
of
lesions
, but
no
recovery
after
other
kinds , and about what " specialization" means in terms of the organization of nervous tissue . This is hard enough to answer with respect
to the primary visual cortex , the somatosensory cortex , the motor
strip , and so forth , but it is far more difficult with regard to the in between areas referred to as " association cortex ." Cowan (1981) gives
a concise summation of the problems :
It is not only that we lack a straightforward
means of activating
most of the associational areas; for many of these areas we hardly
know what types of questions to pose . (p . xviii )
Contributing
and higher
type
of research .
174
voluntary behavior . Sections 5.2 and 5.3 will therefore focus on the
lateralization research, section 5.4 will discuss techniques for determining intrahemispheric specialization of function , and section 5.5
will briefly summarize some neurological data that may help to unseat
assorted philosophical prejudices about higher functions and what
we
know
of them
ing from the epileptic hemisphere to the normal hemisphere . By cutting the corpus callosum, such spread might be curtailed . The
surgical procedure was first attempted by William van Wagenen and
R. Yorke Herren in the 1940s, but without significant clinical success- but also, surprisingly , without significant impairment to the
patient . Apparently Van Wagenen did not usually section the anterior commissure, and he may not always have entirely sectioned the
corpus
callosum
Puzzling over what this large mass of fibers did in the brain , researchersbegan to study in animals the effects of callosal section, and
in the 1950sRonald Myers and Roger Sperry discovered that with the
proper method , they could teach one hemisphere of a cat's brain to
perform a certain task, while the other hemisphere remained ignorant
of it . In their procedure they cut not only the commissures but also
the optic chiasm (figure 5.2), which meant that all information from
the right eye went to the left hemisphere , and vice versa. So merely
by patching one eye, they could exclusively condition one hemisphere to respond to a visual stimulus . This was a remarkable result ,
176
for
Some Elementary
it
of
seemed
to
whether
imply
or
higher
- order
about
why
epilepsy
apparently
Some
two
entire
dozen
of
commissure
patients
was
studies
showed
effects
the
results
the
Dartmouth
done
in
group .
is
Disconnection
This
the
as
In
) .
In
to
1977
anterior
the
referred
as
Technology
Donald
to
as
the
these
the
to
of
but
with
surgery
was
spared
contrast
California
and
Wilson
technique
in
.
,
unrelated
was
,
of
by
definite
method
series
the
set
patients
not
the
as
This
of
1970s
use
the
well
) .
' s
were
his
which
studied
commissure
Eastern
in
ingeniously
the
achieved
to
series
This
the
or
first
the
Bogen
Effect
work
could
clearly
show
the
life
the
two
or
established
definite
or
integrity
effect
condensed
1978
one
one
is
subject
to
on
one
visual
milliseconds
tients
an
one
visual
only
word
other
,
to
,
the
1968
are
posed
hemispheres
can
ensure
the
right
known
( Gazzaniga
) .
to
whether
For
the
example
used
roughly
200
' s
cats
pa
that
flash
these
ordinarily
confining
that
to
the
is
period
means
which
exclusively
see
&
Sperry
. By
speak
1974
sent
and
which
well
Sperry
to
designated
to
5 .2 )
is
tachistoscope
Myers
experimenter
hemisphere
" spoon
chiasma
both
( figure
hemisphere
and
for
( Unlike
optic
goes
hemifield
one
the
5 .3 ) .
intact
information
the
midpoint
hemifield
( figure
have
sual
to
the
up
here
is
so
under
show
Sperry
question
one
activities
procedures
information
cognitive
had
the
to
visual
conditions
the
suffice
and
then
Since
made
- Agresti
available
fixates
signal
be
setup
suitable
hemisphere
.
will
Levy
and
information
own
can
experimental
hemisphere
under
each
its
description
LeDoux
In
between
that
of
disconnection
and
that
disconnection
hemispheres
unity
highly
If
and
,
1965
effects
began
.
produced
cats
al
was
Wagenen
the
Bogen
intracta
The
to
two
Joseph
for
Institute
Van
that
et
Vogel
California
on
School
often
system
sectioned
and
and
the
with
known
which
commissurotomy
obtained
became
serIes
at
the
nervous
success
was
Fisher
moreover
stages
some
questions
procedures
clinical
studies
( Wilson
two
the
matter
of
puzzling
commissurotomy
group
and
not
commissurotomy
callosum
Medical
differences
earlier
tried
intensively
that
,
earlier
some
series
corpus
earlier
was
,
raised
on
underwent
his
the
striking
' s
( Bogen
and
to
it
undoubted
subsequently
Sperry
Contrary
time
the
pair
information
lavished
again
patients
length
cerebral
1960s
this
the
same
and
Wagenen
the
and
anterior
Roger
Van
in
systems
Vogel
of
the
principle
upon
ble
unity
received
processing
Peter
the
they
unifying
Improving
of
that
not
evolution
cognitive
and
Neuroscience
the
the
signal
signal
to
goes
. )
"
is
flashed
to
hemisphere
vi
and
the
RIGHT
"
\
I
\
177
1<'
/
"
,
"
"
I
,
"
I
\
NASAL
TEMPORAL
HEMIRETINA
HEMI
TEMPORAL
HEMIRETINA
RETINA
OPTICNEAVE
OPTICCHASMA
OPTICTRACT
RIGHTVISUALFIELD
LEFTVISUALFIELD
Figure5.2
Informationin the right visual field goesto the left hemisphere
, and informationin the
left visual field to the right hemisphere. Informationpresentedto eacheyeis projected
equally to both hemispheres
. In order to ensurethat informationis presentedto only
one hemisphere
, the subjectmust fixate the midline and the visual stimulusis briefly
flashed. (FromGazzanigaand LeDoux(1978). TheIntegrated
Mind. New York: Plenum.)
subject is asked to report what he saw, he answers " nothing ," but if
he is asked to use his left hand to feel under a cloth for the object
seen, the left hand picks out a spoon. Evidently , the right hemisphere
knows that it saw a spoon and hence the left hand acts on that
information ; the left hemisphere knows nothing and hence, bei~g the
linguistically fluent hemisphere , says it knows nothing .
In another study (Levy , Trevarthen , and Sperry 1972), subjects
were asked to fixate the midpoint , and then a composite (chimeric)
figure was flashed . Generally the subject's verbal reports corresponded to the picture flashed to the left hemisphere, and when the
subject was asked to point to the person seen, the hand (either right
Figure5.3
The basic testing arrangementused for the examinationof lateralizedvisual and
stereognosticfunctions. Here the subjectpicks out by feeling with his left hand the
object that was visually presentedto his right hemisphere. (From Gazzanigaand
LeDoux(1978). TheIntegrated
Mind. New York: Plenum.)
180 SomeElementary
Neuroscience
mind ? How do we know which of these things intact brains have one
of and split -brain subjects have more than one of? What is becoming
evident is that we do not have an adequate theoretical framework
within which we can answer these questions , and that the disconnec -
tion effects should be taken as an occasion for finding out more about
the brain such that we can begin to understand the results, rather
than for forcing the results to fit with our prescientific assumptions .
(The question of the revision of our folk psychological preconceptions
will be discussed at length in chapters 8 and 9.)
survives
brain
death
, this
is a crucial
consideration
. The
brain
may rot entirely , but the soul, argues the dualist , is immortal .
Accordingly , to deny the relevance of anatomical disconnection to
unity of self, dualists have argued with some ingenuity either that
everyone normally has two minds (Puccetti 1981) or that no oneeven the split -brain patients - has more than one (Eccles 1977) . On
in each
case .
The first view entails believing that each skull houses two persons/
two minds , and that even oneself is not , so to speak , oneself . The
selfhood
The trouble is, however , that several split -brain patients have lin guistic capacities not only in the left hemisphere but also show in creasingly impressive capacities for comprehension and expression in
the right hemisphere (see the next subsection). These caseslook like
181
But what conditions must be satisfied for an organism to be con scious , or for an organism to have more than one center of conscious ness? Does the fact that dolphins sleep one hemisphere at a time
mean that they have more than our complement of centers of control /
consciousness /cognition ? Unity of self apparently does not require
complete connectedness of information , since humans characteris tically fail on occasion to notice patent contradictions in their beliefs
and even in their desires (Stich 1978, Nisbett and Ross 1980), so how
much connectedness and integration is required for unity of self ? Are
the split -brain subjects so very different from those humans with split
182 SomeElementary
Neuroscience
personalities ? What is the basis for postulating distinct centers or
modules ? These are empirical questions for which we need empirical ,
as opposed to a priori , answers. (For a further discussion of revisions
to the concept of consciousness, see chapter 9.)
Until much more is known about the mind -brain , these questions
will have to wait for an answer, tantalizing though they may be. As
things stand, the notions of a center of consciousness, or a center of
control , let alone " mind ," " self," '/person," and " soul," are theoreti cally so ill defined that we are at a loss to know how to count such
things . In the absence of a psychological and neuroscientific theory
concerning cognition , ,consciousness, attention , and so on, counting
centers of consciousness and the like is essentially a guess-and-bygolly affair .
Until we know what we are counting , we cannot begin to count and we cannot even say with much confidence that we have one of
whatever it is that split -brain subjects seem to have two of . It is like
trying to count blood types before there was a theory about the constituents of blood and how they differed from organism to organism,
or like trying before Cantor to decide whether there was more than
one infinite set; without a sound set theory , it was like clawing at the
air . To the amusement of future historians , current debates on the
number of selves in split -brain subjects may be seen as akin to
nineteenth -century debates in biology over whether each organism or
organism-part had its own vital spirit , and over what happens to the
vital spirit of a bisected worm whose parts squirm off in separate
directions to begin lives of their own . What I find especially important
in the split -brain results is the suggestion that our familiar, conceptions, such as " center of consciousness" and I'self," do not have the
empirical integrity they are often assumed to have. (For more discussion on these questions, see Patricia S. Churchland 1983, Dennett
1978b, 1979, Griffin 1984, and the commentaries on Puccetti 1981.)
Clues to Lateralization
If the two hemispheres have different capacities, some evidence for
those differences ought to show up in the behavior of split -brain
subjects when the hemispheres are separately stimulated . Hence,
such subjects present a unique opportunity to investigate whether
there are differences and what their nature is. To those of us looking
in from the outside , the early results of studies investigating this
possibility appeared to show that the left hemisphere (LH ) had lin guistic capacities that the right hemisphere (RH) lacked, and that the
RH had certain nonverbal skills in which it was superior to the LH ,
for example, facial recognition and manipulospatial skills .
183
184
the RH lacked expressive capacity for language. The problem therefore was to determine the nature and extent of comprehension by the
RH . Subjects from this series appeared to have severely limited but
significant linguistic comprehension in the RH; the RH could match
simple , concrete nouns with pictures, but performance dropped on
complex nouns and on verbs. When words were spoken aloud , the
left hand could retrieve the matching object. Although it was argued
that the RH must understand language to the extent of understand ing the task demand , this was tendentious
since visual
demon -
RH
never
did . Nevertheless
, the
verbal
command
to make
a fist
wi th the left hand or raise the left arm could be obeyed . Moreover ,
To deny the RH any linguistic facility was too strong, since the left
hand could point to and retrieve objects matching words and vice
versa, and when the RH was asked, " What do monkeys eat a lot of?" ,
the left hand responded by searching in the cloth bag of plastic fruit
and pulling out the banana. In any case it remained a possibility that
the testing methods might favor the LH and give misleading results,
and that the RH might have a greater capacity than hitherto detected
the
California
series
. Given
time
to
scan , the
RH
demonstrated
rather complex comprehension , performing roughly like a normal ten -year old and proving
and
Neuropsychology
andNeurology 185
A
-
Figure 5.5
The Z lens. The lens set-up keeps the patient 's field of view lateralized to one hemisphere. One eye is patched, and the image is projected to only one-half of the retina of
the other eye. (From Springer and Deutsch (1981). Left Brain, Ri(~ht Brail1. W . H .
Freeman and Co. Copyright @ 1981. Adapted from ZaideI1978 .)
186 SomeElementary
Neuroscience
when the request " Smile" was visually presented. Syntactic perfor mance was still inferior , as measured by the Token Test in which
subjects are asked to execute such commands as " Pick up either the
blue circle or the red triangle " and " Put the green circle on the blue
triangle ."
This work greatly complicated the theory that language- motor
control of speech, comprehension , the works- was lateralized to the
left hemisphere . Some researchers came to favor the theory that aspects of linguistic skill are lateralized in varying degrees and that
comprehension is more bilateral than had been thought previously ,
though the left hemisphere is preeminent for motor control . Exclusive
control it probably does not have, since it is known from neurological
cases that aphasic patients with LH lesions can sometimes utter
oaths, sing well -known songs, or repeat well -rehearsed phrases, as in
the case of the proctologist who was severely aphasic but who would
on occasion flawlessly deliver his ritual diatribe against the use of
Preparation H . Allegedly I the " automatic" nature of such utterances
implies that they are more like reflexive responses than like inten tional speech, but matters are too ill defined here for that claim to
mean a great deal.
Zaidel' s results indicated an even greater linguistic capacity of the
RH than the first studies had demonstrated , and one resolution of the
disparity was to claim that the earlier performance profile of the RH
may have been an artifact of the experimental procedure . On the
other hand , Zaidel 's work is by no means beyond controversy . Skepticism over whether the Z lens can be counted on to restrict light
absolutely to one hemifield has inevitably and reasonably been
voiced . Moreover , Zaidel ' s results are based essentially on just two of
the split -brain patients , L .B. and N .G., and in Gazzaniga's earlier
testing of six subjects in the California series, including L .B. and
N .G., the four others apparently failed to show any evidence of RH
language functions (Gazzaniga 1983). Generalizing from the sample
consisting of L .B. and N .G. may well be misleading , therefore, the
more so since studies of the Eastern series of patients revealed that
only three of twenty -eight patients had any linguistic comprehension
in the RH (Gazzaniga 1983). The question of the extent and nature of
the RH's capacity to comprehend language became increasingly complex, and the disagreement began to heat up .
Other complexities began to crowd into the picture . The reliability
of tests in which questions were presented out loud to the subject
came under scrutiny . There seems to be some scope for ipsilateral
(same side) mediation of hand movements in addition to the more
standard contralateral (other side) control . When a command is given
187
out loud , it is received by both hemispheres and the LH may sometimes initiate left hand movements . Accordingly , some of the early
tests showing good linguistic comprehension by the RH may
have been queered by an insufficiently careful method of stimulus
presentation .
To avoid this sort of contamination , Kimura 's (1967) technique using auditory stimuli seems preferable. The anatomical situation is as
follows : in the auditory system there is a major projection from each
ear to the contralateral hemisphere , but there is also a smaller ipsilat eral projection . (See figure 3.7.) Ordinarily , a messagepiped into one
ear goes to both hemispheres . However , Kimura argued that when
there are competing messagesfrom each ear, only the messagein the
contralateral pathway gets through and the message in the smaller
ipsilateral pathway is blocked . In order to deliver an auditory message selectively to one hemisphere , it is necessary to pipe a different
message to each ear. Accordingly , in the testing arrangement the
subject wears headphones and a distinct message is simultaneously
presented to each ear. Kimura 's technique for testing asymmetries
by presenting competing messages to the ears is called " dichotic
listening ."
Although the theory has been generally accepted and the dichotic
listening method has been widely used in studying lateralization in
normal subjects as well as split -brain subjects, disquieting evidence
that the ipsilateral messagedoes sometimes get through has begun to
emerge. Gordon (1980) found in his study of the California series of
split -brain subjects that the commands presented dichotically to the
left ear did sometimes reach the LH - often enough to raise a caution
about results obtained using the method .
As the capacities of the RH were explored in greater detail , it was
found that if the LH of split -brain subjects was occupied in an extraneous task, the RH would respond to a dichotically presented
command such as " Point to the ceiling" with the appropriate action
(Gordon 1980). In one rather dramatic example where different commands were sent to the two hemispheres, the left arm did the bidding
of the RH and pointed to the ceiling while the right arm at the behest
of the LH scratched the table. It might be reasoned that the lack of
response by the RH in earlier tests to the request " Smile" was due not
to failure of comprehension on the part of the RH but to other factors,
such as dominance of the LH for control of brain stem mechanisms
involved in motor ouput .
Clinical studies of patients with right hemisphere lesions give only
ambiguous support to the hypothesis that normally the RH has no
role in language, inasmuch as some studies report deficits whereas
188 SomeElementary
Neuroscience
others can find no significant differences in language performance
between RH lesioned patients and controls . For example, some neurologists (Critchley 1962, Gainotti et ale 1983, Gardner et al. 1983)
have claimed that with sufficiently subtle tests, striking linguistic
deficits can be seen in RH lesioned patients . Such patients apparently
show diminished
spontaneity
ing in finding the right word , and difficulty in giving definitions and
paraphrases; moreover , they have difficulty seeing the point of a joke
or are obtuse with metaphor and double entendre, and they have
trouble summarizing a fable or telling a brief story . (See also Myers
and Linebaugh 1981.) On the other hand , a well -controlled study by
Brookshire and Nicholas (1984) found that RH lesioned patients were
of tumors
of such a case ,
Zollinger (1935) observed that the patient recovered some speech and
was able to carry out verbal commands . More recently , Burklund
and
Smith (1977) report two cases of left hemispherectomy on right handers in which the patient showed a consistent and gradual in -
priately and asked the nurse, " You got a match?" Five months after
surgery he made no errors in counting or in reciting the days of the
week, and only two errors in completing eight sentences. His spontaneous speech was limited , but his residual verbal capacities were
striking .
Clearly , in these patients the RH was not mute , let alone alin guistic , and although neither survived the surgery longer than two
years, the pattern of recovery invites the speculation that perfor -
and Neurology
Neuropsychology
189
190 SomeElementary
Neuroscience
sponses at all to stimuli presented to the RH, not even to simple
perceptual matching tests. Others are capable of some rudimentary
linguistic comprehension , and still others display assorted shades of
proficiency . In addition , the sample size is small, and all split -brain
subjects have a neurological pathology - namely, epilepsy .
Moreover , there is just enough evidence to give life to the suggestion that the greater the linguistic competence of the RH, the greater
is its skill in the other things it is allegedly good at, such as
manipulospatial tasks (Gazzaniga and Smylie 1984). Those who show
no capacity for language even on simple matching tests apparently do
not show RH superiority on constructive and spatial tasks. No response is not very interesting , and consequently the subjects with
nonresponsive RHs are not the ones who are studied intensively . Are
they atypical ? It is not yet clear who is typical .
Two patients from the Eastern series have been found to have
striking and sophisticated RH capacities not only for linguistic comprehension but als~o for expression (Gazzaniga and LeDoux 1978,
Gazzaniga 1983). They are unusual in this regard, and it is believed
that neurological damage to the LH in childhood resulted in a bilat eral representation of language and that they are therefore not typi cal. Nevertheless, there are two respects in which the results from the
studies of the split -brain subjects P.S. and V .P. merit particular reflection . The first concerns how the LH explains and renders coherent
the behavior initiated by the RH, and the second concerns the cognitive capacities of the RH when it is linguistically competent .
How does the LH of a split -brain patient think about the behavior
intended not by it , but by the intentions of the RH to which it had no
access, either before or after the event? The answer is deeply inter esting, for the LH does not appear to go through a conscious process
of seeing the behavior as uncaused by it , wondering how on earth it
happened , and then deliberately fabricating a story to explain it . On
the contrary , it smoothly incorporates the behavior into its scheme of
things , providing a coherent and plausible explanation without conscious puzzlement or mendacity . Consider the following examples.
A picture of a snowy scene is flashed to the RH and a picture of a
chicken's claw is flashed to the LH . An array of pictures is then placed
before the subject, who is to select with each hand a picture that best
matches the flashed picture (figure 5.6). In this setup P.Sis left hand
selected a shovel to go with the snowy scene, and his right hand
selected a chicken's head to match the claw . When asked to explain
the choices, the verbally more fluent hemisphere responded thus :
" That's easy, the chicken claw goes with the chicken and you need a
shovel to clean out the chicken shed." The response of the LH thus
Neuropsychology
and Neurology 191
--- ~-
~ ': 1 -
,~
- - - -
---- ---~. ~l
,
"
'
'
[i11
Figure 5.6
The method used in presenting two different cognitive tasks simultaneously , one to
each hemisphere . The left hemisphere was required to select the match for what it saw
(the chicken claw), while the right was to select the match for what it saw (the snowy
scene). After each hemisphere responded, the subject was asked to explain the behavior . (From Gazzaniga and LeDoux (1978). The IntegratedMind . New York : Plenum .)
makes sense of the left hand action using the resources available to
the LH , and the response is unstrained and coherent.
P.S. understood the surgery , and he understood that his hemispheres were disconnected and could act independently of each
other- or perhaps I should say that his LH and possibly his RH
understood these things . But notice that in the case in question the
LH does not respond by saying that what the left hand does is not its
responsibility , or that the RH controls the left hand, or, most
significantly , that it does not know why the left hand reached for the
shovel. The LH responds as though the motivating factors for the left
hand 's action were its own .
What is striking , therefore, is that given its understanding , the LH
of P.S. does not explain the actions of the left hand as something the
RH did . Even when just reminded of the RH's ability , the LH persists
192
Some Elementary
Neuroscience
in explaining the action as though the action were something it intended all along to do . In caseafter casethe LH integrates the actions
initiated by the RH into its own scheme of things . (For more examples
and discussions, see Gazzaniga 1985.)
This confabulation appears to be natural , untroubled , and appropriate, and there is no evidence whatsoever that the subject is consciously making things up . Other commissurotomized subjects show
that same apparently irresistible tendency to quench ostensible
paradoxes by confabulation . (See Gazzaniga 1983 and 1985.) It is as
though the theoretical framework the brain uses to make sense of its
world will not tolerate or will not accommodate certain kinds of aberrant data or aberrant thoughts . The data that parts of one' s body act
without one's intentions may be consistently rejected by healthy
brains in favor of some more plausible hypothesis . Or thoughts concerning disunity may never earn a working place in a brain ' s representational system of the world because they are inconsistent with its
basic principles for organizing its world , or if they do, the effects on
the representational framework may be a disastrous disintegration .
Making sense of the world is probably as basic a function as just
about anything else the brain does, and the nature of the representational system and the principles of its organization can be investigated empirically (Gardner et ale 1983). Moreover , from an entirely
different area of research, evidence accumulates that normally, explanation of choices, moods, judgments , and so forth , are less introspection -driven than they are theory -driven (Nisbett and Ross 1980).
What seems exciting and promising is that the results from this research on split -brain patients , the results from social psychology , and
the philosophical theory underwriting revisionary materialism (chapters 8 and 9) are converging . (See also Patricia Churchland 1983.)
The first point leads to the second, which is that the real value of
the split -brain research may lie less in what it reveals concerning
anatomical organization - which hemisphere is dominant for which
capacities- than in the light it sheds on the cognitive organizationof
the brain . After all, it could be said that neurological studies (stroke
patients, Wada tests (see section 5.3), and so forth ) had already revealed roughly as much as the split -brain studies about lateralization .
With split -brain patients it is possible to investigate such things as
what difference linguistic capacity makes to the cognitive skills of an
RH (the evidence so far suggests that it doesn't help; see Gazzaniga
and Smylie 1984), whether there is competition between the hemispheres for processing resources (there seems to be; see Holtzman
and Gazzaniga 1982), whether there is paracallosal transfer of infor mation and what its character is (there is, and it seems to be very
193
194
Some Elementary
Neuroscience
The patient may then be asked to hum a melody , make pitch discrimi nations, or something of the kind , as the capacities of the RH are
in vestiga ted .
Findings at the Montreal Neurological Institute , where the test was
developed, show that some 90 percent of right -handers and 70 percent of left -handers have speech lateralized to the LH . The remaining
group is a mixed bag, with some showing speech lateralized to the
RH, and some left -handers showing bilateral motor control for speech
(f{asmussen and Milner 1977). Before generalizing the results to the
population at large, it should be recalled that all these tests were done
on neurological patients , most of whom were epileptic .
With this much description of the Wada test, it might be imagined
that sodium amy tal is the perfect means for exploring the capacities of
the two hemispheres, for at first glance it looks like reversible hemi spherectomy . The appearance is misleading , however , for a variety of
reasons. Although the test has been an invaluable clinical tool and
has yielded precious theoretical data, it is severely limited .
Becausethere are risks with the drug and with the procedures, the
test is kept as short as possible- just long enough to get the piece of
information that makes it worth the patient ' s risk to undergo the test.
Anything else is a bonus . The whole business usually lasts no longer
than two to six minutes , which means that detailed testing of the
subtleties of RH comprehension of language is just not possible.
Coarse-grained questions such as " Which hemisphere is dominant
for motor control of speech?" can be answered, but more fine -grained
questions requiring lengthy tests cannot .
Unacquainted with the hazards of the procedure, I once asked if it
would be possible for me to undergo a left-sided Wada test. My
objective was simple enough ; I wanted to see what it was like to have
an inactive LH and to find out whether , with only the RH active,
there is still awareness, experience, thinking , and so on . I reasoned
that even if my RH was mute , once the amy tal had worn off , the RH
should be able to transfer memories to the LH and the LH could then
report . And of course I expected that " yes-no" questions concerning
my conscious states could be put to me during the procedure- for
example, " Are you aware in the usual way of what is going on?"
As a number of neurologists have explained, the plan was ill conceived. First, the procedure carries risks one would not normally take
simply for the sake of science. Sometimes a piece of plaque is dislodged from the arterial wall and plugs a blood vessel. Second, I was
told I would not get what I was after in any case, because patients
typically have only confused and scanty recollections of the events
during the test, regardless of which hemisphere is suppressed. As
Neuropsychology
and Neurology 195
Bryan Kolb described it , they tend to give confabulated answers to
questions, and they appear unable to recollect even such striking
things as the paralysis to one arm . Even their descriptions of mood
and feelings during the test are at variance with behavior observed by
the researchers.
For example, a person who was weepy and frightened during the
test may recall feeling happy . The explanation for lack of recall is not
evident , though my assumption that a left -sided test would make no
difference to the RH was apparently naive . For some patients the
explanation may be that there was some " leakage" of amy tal to the
other hemisphere . About half the patients show some crossover on
angiography (a method of tracing the vascular system by injecting it
with radio -opaque dye), and it may be that minute amounts " leak"
into the other hemisphere even in the remaining cases.
Accordingly , taking the Wada test may be roughly analogous to
being abruptly made very drunk , with comparable confusion and lack
of recall. Moreover , it may be that even if the LH has some accessto
the information acquired by the RH during the test, the information is
so discordant with the principles of unity that the brain (or the indi vidual hemisphere) uses to make sense of its world that the LH cannot integrate such implications of disunity any more than the LH of
split -brain patients can, and hence it simply denies the discordant
datum and confabulates.
As an index of the complexity in this domain , one other set of
results using the Wada test is important . Kinsbourne (1971) reported
on three severely aphasic right -handers who had some measure of
residual speech. For example, one could utter single words and name
objects, the second could read aloud , and the third could repeat
words spoken to him . In the Wada tests given these patients it was
apparent that motor control for residual speech was now an RH func tion . That is, a right -sided amy tal injection arrested speech and pho nation entirely , but not so a left -sided injection .
Studiesof Normal Brains
In order to counterbalance hypotheses formed in clinical studies,
neuropsychologists tried to devise ways of investigating functional
asymmetries in the brains of neurologically normal subjects. The
problem of finding a suitable technique was formidable , since what
was needed was a way of lateralizing the presentation of a stimulus to
one hemisphere despite the fact that an intact corpus callosum could
be expected to thwart that aim in routine transfer of information from
one hemisphere to the other . However , it was found that in lateralized tachistoscopic presentations (in which the subject fixates on the
196 SomeElementary
Neuroscience
midpoint and a stimulus is flashed for less than 200 msec in one or the
other hemifield ), normal subjects showed performance asymmetries.
In particular , words flashed to the right visual hemifield were more
likely to be identified than words flashed to the left hemifield .
It was reasoned that the hemisphere that received the stimulus
directly had an advantage over the hemisphere that received it via the
corpus callosum and that , when the stimulus was linguistic , the advantage of the left hemisphere would precipitate out as a perfor mance asymmetry . Why the hemisphere receiving the stimulus
directly should have an advantage- assuming it really does- is still
unexplained , but since the performance asymmetry is apparently a
robust phenomenon , the tachistoscopic technique gained currency
and is now widely presumed to reveal asymmetries in brain function .
With the applicability of the technique justified on the basis of
performance asymmetries, further performance asymmetries are assumed to reflect asymmetries in hemispheric function . Words are
flashed to test for asymmetries in linguistic capacities, while pictures
of faces and arrangements of dots and shapes have been used to test
for asymmetries in nonlinguistic performance . On the whole , the
latter have been less reliable in producing asymmetries than the
former .
The dichotic listening configuration has also been widely used with
normal subjects (Kimura 1967). The theory is that items simulta neously presented to the ears go exclusively to the contralateral hemisphere (the smaller ipsilateral channel being suppressed; but see
section 5.2) and that the hemisphere that receives the information
directly somehow has an advantage over the one whose reception is
indirect .
The rationale here is as before. Verbal stimuli presented to the left
ear are less likely to be identified than those presented to the right ;
hence, the stimuli can be presumed to have lateralized effects, almost
as though they were strictly lateralized to one hemisphere . For example, if words are simultaneously presented to both ears, most
right -handed subjects will identify more of those presented to the
right ear than of those presented to the left . This is known as a right
ear advantage (REA). A left ear advantage (LEA) may be obtained for
melodies. Since there are performance asymmetries, direct informa tion must be favored , and if the hemisphere has an expertise for the
directly received information , performance asymmetries will result .
Thus the theory behind the technique .
Lateralized tachistoscopic presentations and dichotic listening tests
are the major means of studying lateralization of function in normal
brains, though other techniques have been used. For example,
197
198 SomeElementary
Neuroscience
sorted studies was impressive (but see McKeever 1979), and a robust
an index
of ?
Finally , there arises the inexorable and vexing question of what the
results mean. Do they show that a capacity is lateralized ? Or merely
that one hemisphere is dominant for that capacity? And what exactly
does " dominant " mean here? Or do they show that one hemisphere
is more efficient than the other in the performance of a certain task?
What precisely are the capacities in question ? Does the size of the
REA reflect the degree of lateralization , or the inefficiency of commissural transfer , or what ? (See Berlin 1977.) Might the performance
asymmetries sometimes reflect idiosyncratic differences in attentional
bias (Kinsbourne 1974, 1982), in listening and viewing strategy
(Springer 1977), or in assorted other variables yet to be factored out ?
How
much
farther
ahead
are
we
if all
we
have
learned
from
these
of literature
reporting
behavioral
asymmetries
with
to escape
Neuropsychology
andNeurology 199
difficult as, say, doing lesion studies in animals, and the equipment is
minimal . As techniques go in neuroscience, they are easy.
On the other hand , it has allowed researchers to simply go asymmetry hunting , and notice that just about any asymmetry is report able. If you are bent on finding an asymmetry , just keep redesigning
the stimulus classes until you get one, for the possibilities here are
quite literally endless. Moreover , as the volume of literature grows ,
asymmetries become even more reportable as they fit or fail to fit with
other reports already in the literature . Trying to determine what is
and what is not significant in the literature surpasses the herculean,
and it may be that , as Stevan Hamad once said to me, the best thing
to do is wait ten years until all the chaff has blown off , and we can see
what is really solid and of lasting value .
In attempts to characterize the deeper capacities underlying the
observed performance asymmetries, it has been argued (Levy-Agresti
and Sperry 1968, Bradshaw and Nettleton 1981) that the hemispheres
use distinct processing strategies or styles and that they use different
codes and different modes of operation . Specifically, it has been
claimed that the LH is an analytic processor, operating in a sequential
manner, and specializes in handling temporal sequencing, whereas
the RH processes in a synthetic , Gestalt, and holistic fashion, using
parallel rather than sequential processing, and specializes in spatial
information . For short, the LH is an analytic-temporal processor, and
the RH a synthetic -spatial processor.
Though the drive for a deeper characterization of capacities is laud able, this hypothesis almost certainly gets us no closer to the truth .
For one thing , there is no theory about what analytic processing is,
and certainly no one has the slightest idea what I'holistic " or IIglobal"
processing might be. Assuming that the LH is specialized for language skills and the RH for facial recognition , it in no way follows
either that the underlying neural mechanisms in the LH are languagelike or that those in the RH are like the processes we are introspectively aware of when we recognize a face. Recognizing a face seems,
to introspection , to be an all-of-a-bunch affair , but from the point of
view of the brain it is hardly likely to be so, considering that the
process must begin at the retina with the excitation of individual cells.
(See also Allen 1983.) Second, dividing space and time across the
hemispheres looks contrived , since many perceptual problems and
motor problems , for example, certainly involve both parameters
(Morgan 1981, McKeever 1981).
There are also nagging methodological problems with the hypothesis. What kinds of tests show that a holistic processor is at
work ? Or an analytic processor? The hypothesis is far too vague to
200 SomeElementary
Neuroscience
inform here, and the worry is that any given performance asymmetry
can be taken as showing either holistic processing or analytic processing , depending on how one chooses to describe the operations under lying it . (Seeespecially Marshall 1981.) The dangers of reasoning from
a performance asymmetry , to a designation of the hemisphere responsible, to styles of processing, and thus in a circle are entirely too
real.
Furthermore , it is surely implausible to suppose that the brain deals
with space and time separately; from what we do know of its operations , there is a spatiotemporal integration . (See chapter 10; also Pellionisz and Llinas 1982.) At this stage, the hypothesis is re~lly a
metaphor in search of a reality to give it substance, and it may be
more misleading than helpful . Evolutionary studies and comparative
neuroanatomy are likely to be crucial in coming to understand the
nature of the deeper capacities and of the underlying neural processes, and how there comes to be a division of labor . Performance
asymmetries in humans , intriguing though they may be, are probably
the outcome of too many complex operations to be very enlightening
for the deeper questions. (For a recent review of the literature and a
discussion of conceptual issues in lateralization studies, see Allen
1983.)
Summary
Although the descriptions of methods in this section have been kept
to cursory outlines , the overarching intent was to act on the principle
that what the data mean cannot be answered without knowing how
the data were obtained . It is important to emphasize the intricacies
and pitfalls of methods in this research because the data and the
interpretations based on the data are only as good as the methods . It
matters tremendously if , for example, certain performance asymmetries turn out to be artifacts of the testing procedures rather than
consequencesof differential capacities of neural regions . Capacities of
neural regions cannot be tested directly but must be inferred from
behavior together with certain crucial assumptions about other
capacities of other neural regions and about where information goes
to and comes from . Should these assumptiQns be less than certain,
the inference is that much more shaky. And of course the vast majority of such assumptions are still themselves in need of confirmation .
What gives a method unimpeachable credentials is a confirmed
theory of brain function that backs it up , but of course at this stage
what we need methods for is to get a theory of brain function . Thus,
progress is made slowly , in bits and pieces, and with continual readjustment , revision , and modification even of the seemingly secure
Neuropsycholog
}r andNeurology 201
and well -established bits . Though I have focused exclusively on lin guistic capacities in this section, similar difficulties and problems
attend hypotheses concerning nonlinguistic capacities, such as
manipulospatial capacities. Worse and more numerous difficulties
beset hypotheses concerning lateralization of emotions and musical
and mathematical capacities. (For good discussions, see Springer and
Deutsch 1981, Kolb and Whishaw 1980, Allen 1983, Bryden 1982.)
5.4
Fritsch (section 4.3), and during the 1940sit was found that it could be
used as a means of charting the cortex for speech areas in patients
about to undergo surgical excision of epileptogenic foci . Using the
technique, surgeons were able to determine before excision what cortical areas were involved with speech and hence what areas to spare .
electrodes directly on the cortical surface. The patient remained comfortable and awake , with just local analgesia for the craniotomy . Us ing weak current , Penfield stimulated selected areas of the brain
while the patient engaged in specified verbal tasks , and observers
areas
of the
cortex
namely in Broca's area, but also in the posterior zones and in Wernicke ' s area , which on current theory was supposed to be dedicated
to language comprehension . There was also a pronounced interfer ence effect in the superior or supplementary motor area. (See figure
202 SomeElementary
Neuroscience
3.4, Brodmann area 6.) Moreover , the zones marked as susceptible to
interference were much more extensive than expected, and the
boundaries of the cortical speech area varied from person to person.
On the basis of this method , the claim that Broca's area was responsible for speech and Wernicke 's area was responsible for comprehension looked far too simple .
More recently , the technique has been refined by George Ojemann
and his colleagues (Ojemann 1983, Mateer 1983), who test for more
than just interference with speech itself , looking for interference with
language-related behavior such as mimicry of orofacial movements,
short-term memory , and phoneme identification . A number of strik ing and highly discrete effects have revealed themselves.
For example, at one particular location stimulation consistently interferes with naming but with no other function , and yet at a site
merely 0.5 centimeters away there is no interference with naming at
all. At some locations only orofacial mimicry is disrupted by stimula tion , at others only short-term memory , at others some combination
of functions , and so on . In one patient who was bilingual in Greek
and English , the discreteness of location of language functions was
especially dramatic , inasmuch as at one site only naming in English
was disrupted , and at a nearby site only naming in Greek was disru pted (figure 5.7).
The locations at which stimulation disrupts a specified function
seem to be stable in a given individual during the test period , and
there is some preliminary evidence that there is considerable stability
across a stretch of time (on the order of months ), as well as some
subtle changes. Locations vary across individuals , so that sites where
naming is disrupted will not be at exactly the same place for each
person. Ojemann has suggested that the cortical sites participating in
particular language functions may be organized in columnar fashion,
much as the somatosensory cortex and the visual cortex are (chapter
3). This work has taken localization hypotheses in an updated and
appealing direction .
Ojemann and his colleagues have also studied interference effects
when specific areas of the thalamus are stimulated and have found
errors in naming , perseveration in misnaming , and subtle effects on
short-term memory . Clearly , subcortical mechanisms cannot be ignored in trying to understand the business of language. Other subcortical areas have also been found to be implicated in language
(Naeser et ale 1982), and the idea that the cortical structures will tell
the whole story is fast losing ground .
203
Figure 5.7
Stimulation mapping of six language-related functions at nine sites on the cortex of a
3D-year-old female, bilingual in English and Greek. Each site stimulated is represented
by a rectangle, and symbols within the rectangle represent significant errors evoked at
that site. Abbreviations : N , naming in English; G, naming ~n Greek; R, reading; VI ,
short-term verbal memory , with stimulation during input to memory ; VS/ memory
with stimulation during presumed time of storage; YO, memory with stimulation at
time of retrieval ; P, phoneme identification ; M , mimicry (If sequences of orofacial
movements; A , site of speech arrest; F/ site of evoked facial movement and sensation.
(From Ojemann 1983.)
NoninvasiveTechniques
In the yarns of science fiction , a Dr . Strangefish has cunningly crafted
a device which , when placed near the head of our hero, will miraculously read off his every fear, plan, and secret thought . Although
nothing as roundly magical as that has been invented , in the last
decade several new devices have made it possible to IIsee" brains and
brain activity without opening the skull . In their reality -bound way ,
some of these devices are every inch a miracle, and their practical and
204
Some Elementary
Neuroscience
theoretical significance has been enormous . They have made it possible to find tumors before surgery, to locate lesions, to determine
whether zones of the brain are differentially active under differing
conditions , to determine abnormalities in specified zones, and so on .
Unlike the fanciful machine of Dr . Strangefish, they do not permit
brain -reading of thoughts and plans, but they can reveal useful
things , such as areas of heightened or depressed activity , which ,
together with information from other fields in neuroscience, can in form localization hypotheses and clinical diagnoses.
The Electroencephalograph
(EEG) The electroencephalograph
is a simple
device first used by Hans Berger in 1929 that records the effects of
putting electrodes on the scalp (figure 5.8). It was found that if the
ouput of recording electrodes was amplified and converted into lines
on a piece of chart paper, certain characteristic wave patterns were
reliably obtained in a range of distinct brain conditions (figure 5.9).
Highly distinctive patterns are obtained during epileptic seizures and
are now diagnostic .
The EEG has been important as a clinical tool in diagnosing such
things as deafness in infants and in establishing a criterion of brain
death (Kelly 1981). It has also been invaluable in investigating what
the brain is up to during sleep- indeed , it has permitted revolution ary discoveries about the nature of the sleeping brain . Using the EEG
Figure 5.8
Diagram of EEG recording system. The scalp electrode records the voltage changes
generated by brain activity I and the electrode on the ear provides the ground . The
signals received from the brain are amplified and displayed on a polygraph or on an
oscilloscope. (From Thompson (1967). Foundationsof PhysiologicalPsychology
. New York :
Harper and Row.)
Neuropsychology
205
and Neurology
EXCITED
~~
.""~ ~
~~ Iw.~
Ji-"i~
'~
~ ~
RELAXED
\lA
.M
/ItlVIwJJv
~J.JVIf
~ ~ (\J"'J\A."t""\J\.J~~AJ
DROW
5Y Jv
\.'-J"J')W
"'~J"J I
ASLEEP
.I""""\vV
#NW
~'V
"lM
~v -,./ '"""..J'v )\ v"V
",,/INW
'1,
I
DEEP SLEEP
'"JI\! \!J\j\.;~'\-"\.V\,/J\
COMA
1 5 EC.
Figure 5.9
Characteristic
variations
50 } lV .
in states of consciousness . (From Penfield
to monitor sleep , researchers have found that during sleep the brain
activity displays a typical sequence of waveform types , with charac teristic durations and regular patterns (figure 5.10) . On the basis of
superficial criteria , sleep appears to be essentially just one kind of
state , and it has often been assumed to be a passive affair , resulting
more from sensory cool -down than neuronal activity . Hence , it was
surprising to find that from the point of view of neuronal activity the
brain ' s sleep state is not a single , uniform type of state but encom passes at least four different sorts of state , as measured by differences
in pattern produced on the EEG , and that sleep is in some manner
actively brought about by neuronal mechanisms .
206
SomeElementary
Neuroscience
B
Figure 5.10
(A ) EEG recordings during different stages of wakefulness and sleep. Each trace shown
represents 30 seconds. The top recording of low voltage fast activity is that of an awake
brain . The next four represent successively deeper stages of non-REM slow-wave
sleep. Stage 1 REM sleep can be distinguished from stage 1 non-REM sleep only by
additional data concerning eye movements and eye muscle contractions . (B) A typical
night ' s pattern of sleep staging in a young adult . The time spent in REM sleep is
represented by a dark bar.. The first REM period is usually short, about 5 to 10 minutes ,
but REM periods tend to lengthen in successivecycles. The amount of stage 2 non-REM
sleep increases until it occupies all of the non -REM period . (Reprinted by permission of
the publisher from D . D . Kelly (1981). Ch . 40 of Principlesof Neural Science
, ed. E. R.
Kandel and J. H . Schwartz, pp . 472- 485. Copyright @ 1981by Elsevier SciencePublishing Co., Inc .)
Neuropsychology
and Neurology 207
reveal anything about the content of a dream; nevertheless, there is
some evidence that the emotional and visual vivacity of dreams is
greater in the later periods .
REM sleep periods appear to alternate in a set fashion with NREM
periods, with the REM duration starting at about 5 to 10 minutes after
sleep onset and increasing with each cycle. Periods of stage 4 sleep
cease about midway in the sleep cycle, and the stage 2 durations
increase. In addition to the systematicity observable within an indi vidual 's sleeping periods , there are also patterns distinctive of developmental stages. For example, infants show the highest percentage of
sleeping time spent in REM sleep (50 percent at birth , 30 to 35 percent
at age 2, and 25 percent about age 10). There is a distinct decline in
percentage of REM sleep in adolescence followed by an increase in
middle age and another sharp decline after age 60. Percentage of
stage 4 sleep declines continuously from infancy , and typically this
stage is entirely absent in old age. Analogous stages of sleep have
been recorded in nonhuman primates and in small rodents and birds ,
stage 4 NREM as well as REM being observable.
Apparently the brain will maintain something like a quota of REM
sleep in the long run . For example, it has been found (Clemes and
Dement 1967) that when normal subjects are deprived of REM sleep,
a REM rebound occurs on subsequent nights . That is, the REM stages
are both more frequent and longer in duration , and the longer the
deprivation , the larger the rebound . REM sleep deprivation does
seem to be correlated with irritability and increased sexual behavior ,
but so far there has been no solid demonstration that anything like
cognitive disintegration occurs as a result of REM sleep deprivation .
Becauseof the frequency with which psychiatric patients complain
of sleep disorders , one line of research undertakes to explore the
possibility that certain psychiatric disorders may be correlated with
abnormalities in the sleep cycle and its characteristic stages. In particular , it has been suggested that abnormal onset of the first REM
period typifies patients with chronic depression (Kupfer and Foster
1972, 1975). There is also evidence that symptoms in depressives
(endogenous but not reactive depressives) may be relieved to some
degree by giving a jolt to the sleep-awake cycle, for example by depriving the patient of REM periods (Pflug and Tolle 1971, Vogel et ale
1975) .
What the brain is up to during REM sleep or during other NREM
stages is not understood , nor is it settled whether dreaming (REMing
and other mentation ) has a functional role in the brain 's economy or
whether it is an incidental by-product of no use in itself . On the
useful side, Roffwarg , Muzio , and Dement (1966) argue that dream-
occurrence
and
the
rebound
effect
of REM - devrived
.
subjects
.
. it
does seem plausible that when the brain goes into the REM stage ,
some point is served, and the brain is not merely idling , doodling , or
" knitting up the ravell 'd sleeve of care." On the strength of tradi tional conceptions, one is inclined think of sleep as loafing , but from
the point of view of the economy of the brain , it is likely no such
thing .
Of importance
to
mechanism has been lesioned show well -integrated movement dur ing REM periods , exhibiting the usual range of stereotypical behavior
including flight , exploration , rage, and grooming .
Jouvet's hypothesis concerning the function of dreaming says that
at least in lower
animals
the brain
rehearses important
motor -
coordination patterns during dreaming , in preparation for real encounters in the waking state. The rehearsal is made safe by the
brain - stem inhibition
of the output
motor
mechanisms
. In contrast
to
the Crick-Mitchison forgetting function for dreaming , Jouvet' s hypothesis envisages that much of the dream pattern is genetically programmed and that consolidation and streamlining of motor control
are the object.
The work on sleep and dreaming in the last thirty years has been
revolutionary in its impact on our conception of the nature of sleep.
Once assumed to be an essentially passive and largely homogeneous
affair, sleep has been shown to have an organization and dynamic
that negates the old conception . Moreover , the research raises ques tions
about
whether
there
are different
kinds
of conscious
states , and
reorganization
or consolidation
does take place during various sleep stages, this raises further questions about the nature and function of consciousness (waking consciousness at least) in cognitive activity generally . Roughly a third of
209
one ' s life is devoted to sleep , and it may well be that important
cognitive work of some variety is carried out by the brain while sensory input and motor output are curtailed . Philosophically speaking ,
the research on sleep and dreaming is of interest both because it is an
instance of the fragmentation and redesign of a folk concept (sleep ing ) as a result of psychological and neuroscientific research and because it bears upon traditional questions concerning what it is to be
conscious . (See also chapters 7 and 9.)
The discussion of the EEG has so far proceeded without raising the
question of the neural basis (the electrogenesis ) of the recorded
waveforms , and this should be raised now . The answer , in brief , is
that the neural basis is not understood . It might be hoped that a given
squiggle on the recording represents the summation of action poten tials of a functionally significant neuronal ensemble , but so far there
are no solid data to confirm the hope . As Bullock , Orkand , and Grin nell (1977) note , in addition to effects from action potentials there may
well be effects from slowly changing potentials due to an assortment
of cellular and extracellular activities .
Moreover , neurons may be grouped so that some field effects are
canceled , some are amplified , some interfere , and so forth . One of the
more thorough critics of certain uses of the technique , C . C . Wood ,
comments (in unpublished work ),
Whether or not activity in a particular group of neurons is evi dent in macropotential recordings depends upon the geometry of
the active cells and cell aggregates , the pattern of synaptic activa tion , the degree of temporal and spatial synchronization , the
spatial relationships between recording electrodes and the active
tissue , and the amount and form of simultaneous activity in
other groups of neurons .
To get even close to determining the electrogenesis of macropoten tials recorded at the scalp one would need , at the very least , to record
field potentials from neural tissue itself . Moreover , one would need
simultaneous , intracellular recordings from large numbers of adjacent
cells , and here the technology peters out . Notice too that whether the
recorded activity is owed , principally or in part , to a functionally
significant neuronal ensemble has certainly not been demonstrated . As
will be evident from the foregoing discussion of sleep and dreaming ,
this does not entail that EEG data can be dismissed as useless , and
certainly their clinical use does not depend on having settled the
question of electrogenesis for a given macropotential . But for assessing the theoretical significance of the recordings it is important to
210
know that what the EEG recordings signify in terms of the anatomical
and physiological substrate is not understood . This issue will become
especially important for evoked-potential research (see below ).
Event-RelatedPotentials (ERPs) Event-related potentials (ERPs; also
known as " evoked potentials " ) are obtained by taking a series of EEG
epochs (for example, 60 or 100 or 500) and averaging them together,
ostensibly to filter out the noise and retain the signal. For example,
the same stimulus , such as a shock at a specified voltage applied to a
fingertip , is presented on many trials to a number of subjects, the
wave patterns are recorded by the EEG, and finally these recordings
are averaged by the computer (figure 5.11). The ERP waves are standardly displayed as an average across a number of trials .
The theoretical interest in ERPs derives from the possibility that
specific waveforms can be correlated with specific types of cognitive
and subcognitive processes taking place in the mind -brain , and that
on the basis of confirmed correlations the nature of the processescan
then be investigated (Kutas and Hillyard 1984). Given this possibility ,
the defining strategy of cognitive psychophysiology is to investigate
cognitive processes through ERP effects.
Conventionally , the ERP recording is divided into the early components (patterns that occur up to 40 msec after the onset of the
stimulus ) and later components (those that occur after the 40 msec
mark). Components are labeled according to time of peak latency
(e.g., 100 msec) and by whether they are positive in polarity (typically
represented as downgoing on the recording ) or negative (typically
upgoing ). Thus, a P300 is a positive , downgoing , wave occurring 300
msec after stimulus presentation , and an N100 is a negative wave
occurring 100 msec after stimulus presentation .
Some ERP results are both striking and robust . For example, there
is a distinct wave that begins to appear about 31 msec after an electrical stimulus has been applied to the finger , and the size of the wave
varies as a function of the size of the voltage (figure 5.12). If the
stimulus is sufficiently large to be consciously experienced, then an
N100 is recorded . It has also been found that where the subject receives the stimulus on both hands, but attendsto only one hand , there
is a clear difference in late components (PlOD, P470).
This difference between early and late components in sensitivity to
attention is robust . The early components are stimulus bound ; they
occur whether or not the subject is attending and are not significantly
modulated even if he is anesthetized . They are therefore called exogenouscomponents
, in contrast to the later components such as the P300,
212
SomeElementary
Neuroscience
v.C
-A
a"9'\~
'),...I~~-'./\.~"'\.I\V
.I"\.r A/"
V"\t~.1"\-~-vv
v,
I' VI hA
.34
l"""T~i2~..,.I~/4/.""\."/\", - b~v./ ~y~ W\!/
""'~V.J~
12
CJ \ l ~
- ] 1"y
20
d \ .~
31
! 46
I~
50
I ~
100
150
200
Figure5.12
Conscioussensationand somatosensory
evokedpotentials(SEP) for faint stimuli. (a)(d) SEPsto an electricpulse(0.1 msecduration) to the third finger of the left hand in a
normal adult. The SEPis recordedfrom the contralateralvarietal
scalD
&
~ focus. Thrpp
hundred (300) trials areaveraged.Thevoltageof the stimuli deliveredat the time of the
verticalline (hollow arrow) is 9 V (a), 11V (b), 12V (c), and 20V (d). Timein msecalong
the abscissa
. Voltagecalibrationof 1 ~m. The blackarrowsidentify the early positive
SEPcomponentsthat disclosea latency shift for faint stimuli (b,c). No SEPcan be
identified for the 9 V stimuli (a). (e) Detectionby the subjectin the sameexperiment
with a plot of the numberof trials detectedin eachrun (ordinate, percent) asa function
of the voltage of the finger stimuli (abscissa
, volts). The letters identify the correspondingSEPrecordsin (a)- (d). (f ) SEPto near-thresholdelectricstimuli to the index
finger of one hand in a selectiveattentionparadigminvolving randomseriesof stimuli
deliveredto 4 different fingersat a fastrate(meanrandomfrequency150/min). Thetwo
superimposedtracescompareSEPsto identicnlstimuli that were designatedeither as
targetsto be mentallycounted(thickerline) or in other runs asnontargetsto beignored
while the subjectidentified stimuli to the oppositehand (thinner line). The early P50
SEPcomponentis not modified, but large P100and P470componentsoccur for the
targets. (FromJ. E. Desmedt, J. Debecker
, and D. Robertson1979. In CognitiveCompo
nentsin Cerebral
Event-Related
Potentials
andSelective
Attention, ed. J. E. Desmedt. Basel:
Karger.)
required to count objects from one category. It has been found that
large-amplitude P300 waves reliably occur on those occasions when
subjects are presented with the less probable stimulus , and the rarer
the stimulus , the larger the P300 amplitude (Duncan-Johnson and
Donchin 1977). The oddball paradigm has many versions, and the
categories can be quite abstract- for example, male vs. female names,
politicians vs. entertainers , and so forth (Kutas, McCarthy , and Donchin 1977). The omission of an expected stimulus will also yield a
P300 component (Picton et al. 1974). It has been found in addition
that semantically strange sentences, such as " The boy ate eggs and
213
214 SomeElementary
.,I Neuroscience
neuronal processes that obtain from occasion to occasion. A major
concern, therefore, is that if we do not understand the anatomical and
physiological substrate for the ERP, the assumption that the averaging technique extracts signal out of noise may be mere wishful think ing . (For an excellent and detailed discussion of the problems
connected with inferring information about the nervous system from
ERP measurements, see Wood and Allison 1981.)
The nature of the neuronal substrate is rather better understood in
the case of the early components, where the evidence shows them to
be linked to signal transmission from the periphery , and animal studies and lesion studies have yielded converging results. But for the
endogenous components , such as the P300, no one has a well supported theory concerning what the neuronal substrate is. It cannot even be assumed that the effects are owed primarily to cortical
activity , since thalamic activity , for example, could also affect the
recording . Direct recording from the cortical surface could at least
narrow down the alternatives here, but it still will not reveal the
source of the effect.
Assuming that the methodological justification of evoked-potential
research is that it helps bridge the gap between neuroscience at the
micro level and psychology at the macro level, then it must address
the matter of the neuronal substrate for its data. Unless it does so, the
possibility remains very much alive that a given peak or trough really
does not mean anything in terms of a distinct step in the processing of
information executed by a discrete neuronal ensemble.
If there are difficulties in hitching endogenous ERPs to the
neuronal substrate, there are also difficulties in hitching them to in formation -processing functions at the macro level . Already mentioned is the problem of resolving the conditions for eliciting a
waveform so that the waveform can be plausibly correlated with a
specific cognitive or subcognitive process. An additional and serious
difficulty concerns the individuation of the waveforms themselves.
That is, how does one tell when the recording shows a P300 or a
P500? It appears to me that the individuation of componentswhether there is just one, and if so, which one it is- is by no means
settled on the basis of the physical features of the recordings, and not
infrequently involves a generous dollop of interpretation .
The temporal definition for a given waveform is not very fine, and
the timing turns out to be a somewhat flexible criterion for individua tion . For example, Donchin et al. (1983) argue that the P300can occur
anywhere between about 250 and 900 msec and can even occur after
the motor response. The claim is that the onset of the P300 is later
when the discrimination task is more difficult or complex. However ,
215
216 SomeElementary
Neuroscience
really was a P300, but it happened at 800 msec, or it had a low
amplitude , or the conditions of production were not standard, or
there were overlapping components , and so on; alternatively , one
might say there really was no P300because the conditions of produc tion were nonstandard , or the shape was not right , or the time was
not right , and so forth . I do not say these disputes typify ERP research, but they are sometimes in evidence, and it is difficult to see
how such disagreements can be resolved in principled fashion in the
absence of independent criteria for individuation .
Despite its methodological tribulations , ERP research has an
impressive following , largely , as its proponents rightly point out ,
because it is the only noninvasive way of collecting data on the .brain
activity of normal , awake, behaving humans . To be sure, a technique
satisfying those specifications is highly desirable, and consequently
there have been systematic attempts to refine experimental procedures and to construct more revealing mathematical analyses of the
data. Moreover , ERP data can be useful in conjunction with results
using other techniques in psychology and neuroscience, even though
some of the methodological issues are unsolved . Perhaps the best
example of this is the research of Hillyard and his colleagues concerning directed attention .
The basic ERP fact that Hillyard exploits is that the early components appear to be insensitive to directed attention , whereas later
components are sensitive . The N100 appears to be the first component whose variation reflects what the subject is attending to, and it is
modality indifferent . Hillyard ' s insight is that data concerning the
N100- for example, what stimulus parameters it is sensitive to,
whether specific combinations of properties have the same N100
profile - might be brought to bear in adjudicating the debate between
early selection and late selection theories in cognitive psychology . For
this aim, it can be left open what unitary function , if any, the N100
manifests .
Roughly , those favoring early selection (e.g., Broadbent 1970,
Treisman 1969) argue that there is some early selection of stimuli
based on simple physical properties and that only selected stimuli
receive full analysis in subsequent processing. On these models the
early stimulus selection is later followed by a responseselection. By
contrast, in late-selection theories (e.g., Norman 1969) the stimuli are
thought to be fully analyzed before any selection occurs. Since the
available behavioral data do not decisively cut against either theory ,
Hillyard and his colleagues have tried to design experiments that will
provide ERP data indicating whether or not early selection occurs.
The idea in one experiment (Hansen and Hillyard 1983) is to mea-
217
sure how much difference it makes to the N100 whether a two property stimulus has the right combination of properties to make it a
candidate for a target stimulus . Let the stimulus be sound pips, and
the properties be pitch and location . Subjects are to attend , for example, to high -pitched sounds emitted from the left; of sounds in this
class, the target sounds will be those with long duration . Put crudely ,
the question is: if a sound has the " wrong " location , does that affect
whether or not the other property , pitch , is processed? Hansen and
Hillyard ' s analysis of the NIOOindicates that it does affect whether or
not there is NIOO sensitivity to pitch in such a case, suggesting that
the properties are not independently and fully analyzed, as the lateselection theory advocates, but that only selected stimuli are fully
processed. (See also Hillyard et al. 1984; for a review of ERP research
and cognitive theories, see Kutas and Hillyard 1984.)
5.5 ImagingTechniques
ComputerizedAxial Tomography(CT Scanners
)
The principle of computerized axial tomography (CT scanning) is to
rotate around the head an X-ray source that is coupled to a synchronously moving X-ray detector on the other side, and then to use a
computer to reconstruct in visual form the variations in X-ray opacity
of the brain . This gives a cross-sectional X-ray picture of the brain at
one level or cut (hence " tomography " ). For other " cuts" the procedure can be repeated at a different level . Normal brain scans are then
compared to the scan under study to determine whether there are
abnormalities .
This system was the pioneer of the computer -aided visual displays
of the brain , and although it has been a valuable innovation , it has
considerable limitations even as a diagnostic instrument . If an abnormality such as a tumor is to show up , it must have absorption proper ties different from those of normal tissue, and this is not always
. the
case. Abnormalities may not show up on a CT scan until they are in
the serious stages; small tumors may be missed, and demyelinating
diseases such as multiple sclerosis are missed more often than they
are detected. Additionally , there is always some risk in exposing the
body to ionizing radiation, and frequent exposure is to be avoided. On
the other hand , CT scanners are simple to use and are much cheaper
than PET scanners (see below ).
CerebralBlood Flow (rCBF) Studies
This technique exploits the assumption that blood flow in a cerebral
region is coupled to the metabolic needs of neurons in that region ,
218
Some Elementary
Neuroscience
Neuropsychology
and Neurology 221
anatomy . The limits of the technique have certainly not been reached.
The major drawback to the PET scan technique is expense. It requires
a powerful computer , a cyclotron 'to manufacture the needed isotopes, and technicians to run the machines.
Nuclear Magnetic Resonance(NMR) Imaging
Nuclear magnetic resonance (NMR ) imaging produces far better im ages than the CT scanner, and without the use of radioactive substances and ionizing radiation . The secret of NMR imaging depends
on a curious property of the nuclei of atoms: if the nucleus contains
an uneven number of particles, then it has a spin and hence behaves
like a magnet. Hydrogen and phosphorus , of which the body has
large amounts , are atoms with such nuclei . When a substance containing such nuclei is placed in a magnetic field , the axis of spin of the
nuclei is altered, and the magnets/nuclei orient themselves in confor mity with the lines of force. Once regimented , they can be made to
point in a new direction and then allowed to flip back. This manipula tion provides information , for as the nuclei flip back, they release
energy, and this emission can be monitored and used to reconstruct
an image of where the hydrogen , for example, is concentrated and
distributed .
The idea is to exploit the fact that substances such as hydrogen and
phosphorus have nuclei that can be manipulated , can be manipulated
selectively, and are unevenly distributed in the body . Hydrogen is of
course found in water and in lipids , and there are subtle differences
in the distribution of hydrogen in soft tissue that show up in NMR
images. Hence, differences between normal brains and brains with
tumors or other lesions are discernible . This sort of image provides
anatomical information , but metabolic portraits - of special interest to
the theoretician - may be obtainable as well . Phosphorus is a part of
adenosine triphosphate , or A TP, which is the energy-storing
molecule providing energy for all metabolic processes, including
those in neurons . A TP is concentrated in regions of greatest metabolic
activity , so greater concentrations of phosphorus can provide clues
concerning what parts of the nervous system are subserving particu lar functions .
The technology for NMR imaging is still in its infancy , and NMR
enthusiasts predict that it will replace CT scanning for diagnostic
purposes and that as refinements are made it will be of increasing
value to neurophysiology and neuropsychology . (For a good in troduction , see R. Edelman 1984.)
222
223
224
may identify his wife by her glassesor a hat- though if these are put
on someone else, he will pick out that individual as his wife .
Recently it has been discovered that patients who are unable to
recognize faces nonetheless show a large electrodermal skin conductance when shown faces familiar to them before the dysfunction , and
they do not show a similar response to unfamiliar faces (Tranel and
Oamasio 1985). In one case the patient also showed the skin response
to faces newly encountered . The data suggest that the physiological
basis for facial recogniti In is intact in these patients , even though
they have no conscious accessto the information .
In part what these cases show is that recognition of individual
conspecifics is not just an aspect of the general cognitive business
subserving recognition of teacups, dogs, trees, and so forth . Something special is required , and thus important questions are raised
about what that is and about the evolutionary significance of special
mechanisms for recognition of conspecific individuals . Since the ability to identify others as individuals is found in many other species,
including some insects (e.g., crickets), the questions raised here connect to questions regarding the brain 's representational scheme of social relations and how it evolved .
V isual Agnosia
Patients otherwise unimpaired visually may be unable to identify by
vision common objects, though they can name them if allowed to
touch , hear, or smell them . They can faithfully copy a line drawing
but remain unable to say what the object is. (See figure 5.16.) Sometimes this syndrome is found with color agnosia (inability to identify
colors) and facial agnosia, and sometimes with alexia, and the lesions
are usually in the areas of the visual cortex lying outside areas 17 and
18. Data from these casesare useful in determining the character and
structure of sensory modules , in hypothesizing what precedes
and follows certain elementary processes, and so forth . (For such a
use, see Marr 1982.)
Blindsight
Anatomical studies of the visual system indicate that the striate cortex
(primary visual cortex) receives a massive projection of neurons from
the lateral geniculate nucleus, and earlier this area was widely regarded as the first major processing region for visual information .
Lesioning the striate cortex was thought to result in complete blind ness, and there was substantial evidence for this view . Nonetheless,
it was found (Weiskrantz and Cowey 1967) that monkeys whose
striate cortex had been excised showed a surprising residual visual
225
)
,
' \
Figure 5.16
Copies of line drawings by a patient with visual agnosia. Before drawing , the patient
could not identify any items, and after making copies, his attempts to identify what he
had drawn were as follows : (A ) " I still don ' t know ." (8) " Could be a dog or any other
animal ." (C) " Could be a beach stump ." (0 ) " A wagon or a car of some kind . The larger
vehicle is being pulled by the smaller one." (From Rubens and Benson 1971.)
226 SomeElementary
Neuroscience
capacity
. They
criminate
tex
was
that
could
two
removed
in
must
to
after
through
which
the
the
brain
split
- brain
superior
, it
colliculus
make
was
could
dis -
striate
cor -
had
suggested
pathway
) . Thus
, to
, there
explain
hypothesized
provided
visual
entire
subjects
1970
lesions
and
the
visual
( Trevarthen
to
obstacles
when
geniculostriate
striate
relied
large
, even
on
major
pathway
capacity
route
work
the
a second
residual
around
patterns
. Other
addition
be
maneuver
- dimensional
the
the
that
the
information
discriminations
on
( figures
3 . 10 ,
3 . 11 ) .
Naturally
comparable
damage
ing
the
. In
removal
of
of
both
run
the
eyes
of
blind
in
his
When
see
answers
with
left
for
to
too
small
, or
the
650
msec
, he
could
of
light
and
where
racy
no
in
of
mere
he
was
indeed
features
This
first
to
duration
trols
have
visual
of
has
field
that
residual
visual
were
strikingly
visual
fields
triangles
and
to
and
circles
pregeniculate
of
was
make
not
that
he
could
shorter
great
had
been
that
his
use
of
did
about
vertical
bars
accuracy
shining
to
( accu
" guessing
visual
not
so
not
than
point
with
do
were
from
he
he
his
discriminations
horizontal
light
that
subject
has
findings
, and
eliminate
the
they
in
visual
" was
information
have
( chiasmatic
followed
consciously
of
were
possibility
to
visual
equiluminant
, two
)
subjects
lesions
that
six
to
the
in
unable
him
blindness
to
had
in
their
patterns
blind
who
O . B . Like
targets
that
con
spared
, and
whose
were
the
subjects
hemisphere
similar
others
and
presented
) tested
one
to
by
procedures
stimuli
( 1978
pointing
contrast
been
testing
the
distinguishing
. By
the
guess
presentations
making
fact
decortication
in
is , on
to
, and
the
incontestable
J eannerod
accurate
field
to
capacities
to
presentation
a point
the
to
complete
as
also
he
early
and
hemi
visual
, O . B . was
merely
environment
could
a human
access
. Perenin
undergone
the
field
refined
right
left
, O . B . protested
able
long
etc . ) . It was
of
the
the
surgical
description
asked
discriminate
O ' s . He
the
been
field
visual
So
correctly
study
confirm
of his
and
concern
" ) . 3 That
own
capacity
tests
a judgment
, despite
visual
information
tend
owed
visual
therefore
guessing
on
in
hemianopia
was
visual
findings
cortex
striate
undergone
blind
his
whether
with
accuracy
blind
their
was
by
of
humans
, had
striate
, he
he
X ' s from
his
the
, and
residual
29 / 30 , 30 / 30
to form
available
field
coarse
impressive
published
, and
these
respect
with
visual
in
reasons
of
tests
question
found
a " homonymous
nothing
. In
ale
surgery
clinical
tested
could
the
the
be
medical
portion
of
( called
standard
et
, for
major
a result
raised
could
Weiskrantz
O . B . who
. As
studies
capacity
1974
a patient
sphere
monkey
residual
their
, they
blind
such
as
was
perform
227
information
anywhere, and thus both the major geniculostriate pathway and the
hypothesized second pathway would be deprived .
It should be mentioned that the second pathway hypothesis is still
contentious , and some researchers believe that the blind sight phe -
of where the point of light is located, yet they can locate it . Sometimes
the subjects acknowledge that they have a " feeling," such as a prick ling or a feeling of " gunfire at a distance" (Richards 1973), but these
experiences are not visual experiences . What is remarkable is that
subjects who deny visual awareness can yet be consistently correct in
blind
It has
visual
sometimes
field
seemed
safe
to assume
that
behavior
at this
level
sidering data just from the blind sight studies, we may not be clear
about which assumption
the cut and dried affair one might have hoped it to be (Swets, Tanner,
and Birdsall 1964). From the point of view of philosophy , it is impor tant to see that this is an instance where empirical discoveries put
228
229
ef-
with
lesions
in
the
limbic
structures
and
does
not
occur
personality of
the patient may also be a factor, though this does not imply that the
syndrome is psychiatric . (See especially Weinstein and Kahn 1955;
Kinsboume
and Warrington
1963b.)
These cases are arresting because they challenge the general assum ptions that if one cannot see or cannot move an arm , then one is
aware
and
aware
without
reservations
that
one
cannot
see or can -
not move the arm , and that if one is making unintelligible conversa tion , then one is aware , instantly , that one is not making sense . Such
assumptions now appear problematic , and thus the blindness denial
and other denial syndromes have implications that suggest revisions
to the commonsense conception of what it is to be aware . Taken
together with other cases that bear upon awareness , such as discon -
nection, hemineglect (see below ), and blind sight, this case makes us
stand back from the commonsense conceptions concerning con sciousness and suggests that there is much to be learned about the
the
behavioral
data , the
usual
clinical
characterization
of
patients suffering from blindness denial is this : they are blind , but
they are not aware that they are blind . However , it has been suggested to me by several philosophers that this case must be misdescribed, because, it is said, the following is a necessary feature of the
concept of consciousness: so long as one is conscious, and so long as
one can make judgments at all, then if one is not having visual experiences, one will be aware that one is not having visual experiences,
and
blind
vice
versa
. I do
inasmuch
of
. In a word
course
, you
concede
cannot
that
as we do not understand
fail
the
to know
case
may
whether
be
the empirical
you
are
misdescribed
situation
well
enough to be certain what the correct description is. However , I cannot agree that we can know a priori that it must be misdescribed . On
230
Neuropsychology
and Neurology 231
~
rt '!>fr
c6
;:(
S
q
{o
, ~ ..-11
~ -- -
Figure 5 .17
(A ) Examiner ' s drawing of a daisy on left ; patient with hemineglect drew daisy on
right . (From Heilman et al . (1983) . In Localization in neuropsychology , ed . A Kertesz . New
York : A {:ademic Press .) (B) Drawings of patients with hemineglect when asked to fill in
the clock faces . (From Bisiach et al . 1981.)
lation
of deficits
present
with
sion
use
and
aphasic
RH
lesions
capacities
their
The
left
. But
happens
what
hemineglect
hemispace
for
is complex
to know
tions
, and
may
be introduced
reference
at the
point
be made
that
LH
into
signing
the
and
to the
signing
patients
? Are
. As
has a crucial
level
discourse
to these
last
and
plane
points
and
such
deficits
patients
able
later
. Relations
to
, it is im at the
level , to make
case , for
associated
, and
is , they
comprehen
a preliminary
in signing
morphological
. In the
horizontal
role
. That
visuospatial
purposes
surprising
signs , at the
syntactic
patients
lesions
communicative
and
space
in the
by
as do speaking
with
level , to differentiate
will
sites
of left
answer
portant
lesion
symptoms
example
with
lexical
inflec
, nouns
an arbitrary
pronoun
between
reference
objects
232 SomeElementary
Neuroscience
specified
by
the
the
spatial
lished
Consider
Poizner
this
is
the
quite
of
Klima
for
"
for
of
normal
tion
It
the
appears
is
relations
cated
in
right
is
points
signing
of
and
hemifield
And
her
com
hemifields
syntactic
space
when
the
left
she
When
However
her
both
for
of
but
in
use
use
used
aware
room
by
room
hemifield
side
to
space
of
estab
tested
her
left
continued
sense
subjects
the
making
when
terms
the
describe
her
omitted
she
in
between
to
in
she
some
specified
purposes
describing
be
asked
signers
between
as
will
signing
When
by
in
furniture
"
that
subject
spatial
rectly
the
movement
hemineglect
syntactic
signs
"
of
1979
result
normally
her
the
square
surprising
prehension
was
representa
and
the
space
is
appropriate
represented
representation
of
di
space
is
trun
One
leading
theory
orienting
mechanisms
nisms
and
and
of
function
hemineglect
or
body
emphasize
subjective
merest
of
view
are
in
Calculating
so
poorly
our
consciousness
the
no
the
theories
mechanisms
about
the
must
for
body
scheme
have
final
word
And
they
and
but
understand
attention
the
phenomena
brain
the
to
and
repre
since
access
means
Bisiach
brain
of
that
event
phenomenological
attempt
by
related
about
that
out
any
with
and
by
In
defect
mecha
understood
concert
arousal
suggests
consciousness
point
in
Klima
questions
the
beginning
is
In
is
defect
carried
and
that
more
point
awareness
representation
once
involved
and
hemineglect
is
research
Poizner
attention
space
that
it
crucially
important
awareness
in
is
the
beginning
raises
attention
be
of
that
although
may
only
is
Bellugi
space
and
constitute
by
1979
another
1977
Luzzatti
nature
Heilman
Kinsbourne
sentations
the
rather
the
the
nature
of
Oddities
The
ability
of
lations
to
ordinary
humans
solve
is
whereby
ing
in
prodigies
savants
1985
his
can
show
their
arithmetic
the
like
great
be
calculating
arithmet
Sometimes
or
may
remarkable
complicated
speed
symptoms
abilities
calcu
is
of
perform
with
autistic
and
perform
and
development
individual
and
to
exceptional
head
also
idiots
an
an
calculations
count
equations
~there
capacities
to
complicated
Occasionally
calculat
may
be
highly
demented
confined
Sacks
Critchley
who
had
Fleury
root
one
all
sign
chased
direction
Bellugi
and
together
the
He
and
squashed
"
the
Klima
now
asked
and
points
Bellugi
ical
verb
relation
of
1979
describes
several
superlative
calculating
described
465
as
484
375
"
destructive
775
severely
abilities
imbecile
in
thirteen
retarded
seconds
"
individuals
famous
who
example
was
produced
and
had
the
prodigious
cube
for
dates , but
was
unable
to
grasp
even
the
233
rudiments
of
geometry . Some such prodigies remember all the stock market quotations from the newspaper
ingenious
compensatory
himself very wealthy . All too aware of his inabilities , he hired accountants to perform the calculations he required , and becausehe saw the
dangers in trusting
done by
there
may
be subtle
connected
deficits
that
are not
detectable
234 SomeElementary
Neuroscience
There
are
the
legions
brain
of
does
discovery
that
have
the
brain
are
mirror
to
around
ond
first
of
the
spatial
and
'
Neurological
the
advent
of
valuable
and
perhaps
over
as
mind
does
can
correlate
in
we
outset
to
intuitions
and
'
and
if
interrupt
More
also
of
new
the
vitally
nature
for
up
course
intuitions
the
room
stokes
the
in
to
our
about
leave
curiosity
With
of
are
unseat
verities
wonder
disease
studies
commonplace
Unseated
about
that
and
repre
neurotransmitter
set
neurological
areas
characterize
unsettle
of
Alzheimer
certain
can
hand
headway
is
of
that
to
then
make
and
'
1980
to
linked
that
left
sec
nature
absorbing
degeneration
1982
to
our
what
theory
theory
and
making
Conclusions
None
of
the
methods
means
for
telling
carried
out
in
fect
and
and
what
dures
be
we
the
brain
al
the
the
possible
if
is
the
the
machinery
with
ravages
our
third
sources
deficits
Perry
at
electrifies
is
read
signature
on
earliest
in
only
his
about
et
to
it
can
to
with
continue
changes
of
Heilman
the
capacity
and
connect
of
that
he
us
can
experiences
drive
tell
See
example
the
their
that
For
and
remarked
assumptions
one
Perry
for
degeneration
reverse
important
they
dementia
the
deficiency
it
space
it
plate
to
case
none
and
visual
that
home
performance
senile
and
with
discover
does
in
is
image
attempts
techniques
lesions
disease
the
to
with
None
is
studies
physiological
of
anatomical
data
information
each
in
its
distinct
common
lesion
behavioral
own
is
from
ERPs
results
useful
Accordingly
studies
recording
data
per
proce
with
fire
are
way
animal
and
is
hypothesis
data
stimulation
sure
functions
experimental
with
higher
procedures
from
conjoined
normals
yields
certain
experimental
be
chapter
. lc
the
upon
to
investigations
with
of
that
converge
need
this
whe
Nevertheless
hoped
slowly
studies
studies
and
flaws
be
in
how
brain
of
is
will
discussed
us
free
human
these
imaging
immensely
were
brain
new
correlating
we
How
body
and
such
cavity
range
'
so
the
ventricles
cranial
What
mirror
from
what
explain
the
normal
patient
the
does
studies
regarding
the
the
we
whom
the
writes
again
one
1985
diamond
What
of
stroke
of
can
above
experiences
representation
of
percent
Berker
normal
mirror
home
understanding
How
in
and
of
baseball
road
sentation
90
result
the
to
side
of
up
normal
degree
as
image
held
runs
the
university
our
organized
over
in
when
books
challenge
is
patients
occupy
IQs
complete
it
hydrocephalic
to
have
that
how
some
expanded
theless
cases
and
electro
and
clinical
both
PET
of
scans
Neuropsychology
and Neurology 235
tachistoscopic and reaction time studies, etc.), neuroembryological
data, and so forth . Whitman Richards (1983:461) sums it up thus :
If so many psychophysical experiments are flawed , why bother ?
One reason that motivates many of us to study behavior with
imperfect designs is that our imperfections will usually differ
from one set of experiments to another, thus allowing a convergence to the understanding of underlying mechanisms.
Selected
Readings
Bryden, M. P. (1982). Laterality
: Functionalasymmetryin the intact brain. New York:
AcademicPress.
Feindel, W., R. S. J. Frackowiak, D. Gadian, P. L. Magistretti, and M. R. Zalutsky, eds.
(1985
). Brainmetabolismand imaging. Discussions
in Neurosciences
2/2 (specialissue).
Galambos
, Robert, and StevenA. Hillyard, eds. (1981). Electrophysiological
approaches
to
humancognitiveprocessing
. (Basedon a work sessionof the Neurosciences
Research
Program, 1981.) Neurosciences
Research
ProgramBulletin20:141- 265.
Gazzaniga
, MichaelS., and JosephE. LeDoux(1978). Theintegrated
mind. New York:
Plenum.
Hecaen, Henri, and Martin L. Albert, eds. (1978). Humanneuropsychology
. New York:
Wiley.
Kertesz, Andrew, ed. (1983). Localizationin neuropsychology
. New York: Academic
Press.
Kolb, Bryan, and Ian Q. Whishaw (1980). Fundamentals
of humanneuropsychology
. San
Francisco
: W. H. Freeman.
Mendelson, WallaceB., J. ChristianGillin, and RichardJedWyatt (1977
). Humansleep
andits disorders
. New York: Plenum.
Perecman
, Ellen, ed. (1983). Cognitiveprocessing
in theright hemisphere
. New York: Academic Press.
Squire, Larry R., and NelsonButters, eds. (1984). Neuropsychology
ofmemory
. New York:
Guilford.
Thompson, RichardF. (1975). Introductionto physiological
psychology
. New York: Harper
and Row.
Chapter 6
Introduction
6.1 I ntroduction
Within the domain of neuroscience, questions at one level of generality inevjtably provoke questions at both higher and lower levels of
generality . Wondering how the retina works , for example, lead~,
down one path to inquiries about the microstructure of rods and
cones, and thus to biochemistry , and eventually to physics. It also
leads in the opposite direction , to inquiries about such things as the
organization of neurons in the visual cortex and what functions are
there carried out . Following that ball of yarn as it unwinds before us,
questions of greater generality take form . How does the brain visually
recognize shapes and colors, how does it recognize crows and water ?
How do we see? Conviction concerning how to go about answering
these questions shades off as the questions become more broad, more
general, or more abstract. As the ball of yarn rolls on, our curiosity
begins to limn the shape of large and ominously general questions
looming out of the fog . Can we develop a science of animal behavior ,
human behavior included ? What will be the role of neuroscience in
such a science? How can neuroscience and psychology be integrated ?
The shape and character of a research program can be molded by
the perspective one has on questions at levels up and down from
where one' s day to day efforts are directed . Neuroscientists , therefore, like scientists in all fields , are compelled on occasion to follow
the ball of yarn past the gates of their paradigm , and thus to contemplate the coherence and consilience of their research within a broader
framework . They are compelled , that is, to be philosophical .
The dawn of natural philosophy , which is to say, of science, was
marked by a fascination with the disparities between appearance and
reality ; disparities between what seemed, given one' s beliefs and
240
some observations, to be true, but what , given other beliefs and observations, seemed to be false. Early philosophers were moved to
suspect that how things appear may be the outcome of a deeper
reality , hidden from view , and radically different from the appearance apprehended in observation . Their reflections on the nature of
substance, change, motion , fire , and so forth , gave rise to bold theories about the deeper reality and about why it was not immediately
perceived. Thus, for example, in a swashbuckling slice through the
blooming , buzzing , confusing complexity the world presents, the ancient Greek philosopher Democritus (460- 370 B.C.) claimed that the
reality underlying appearances consisted of nothing but atoms and
the void . The atoms, he hypothesized , hooked together in assorted
ways, and differences in the appearance of substances were owed to
differences in the shapes and collective organization of the funda mental atoms. Humans , like everything else in the universe, were at
bottom just organized assemblie~ of atoms, although the organization
was neither understood nor assumed to be simple . Minds , therefore,
were conceived as being fundamentally material , and their remarkable capacities were a function of the remarkable organization of
matter .
The force and vivacity of these ancient ruminations can be recaptured by recollections of appearance/reality discrepancies in one's
childhood . Recall watching the moon scudding by the clouds on a
windy night , and having that seemingly unassailable observation corrected with the claim that it is the clouds that are doing the scudding .
How do you know ? Line up the position of the moon with the top of
the poplar tree, watch for three minutes , and see what has moved . A
new observation conflicts with the old .
Recall the surprise at being told that the earth is a huge sphere and
that it moves around the sun . How do you know ? The answer here is
longer , requiring gravitational theory and Newton ' s mechanics, and
of these in turn one can ask, How do you know ? In the third century
B.C. Aristarchus offered a heliocentric view ; the Sun was at the center,
and the planets, including the Earth, moved about the sun in circles.
The rationale for this theory derived from calculations Aristarchus
had made of the distance of the Moon and Sun from the Earth, and on
that basis he roughly gauged the relative sizes. Since the Sun was so
vastly larger than the Earth, he thought its daily passage across the
sky must be mere appearance, and that in reality it was the Earth that
moved rather than the Sun.
Although his calculations were acceptable, Aristarchus could not
explain how the Earth moved , since nothing was pushing it along,
and prevailing Aristotelian theory requ~ed continuous application of
Introduction
and History
241
6.2 EarlyEpistemology
Plato (429- 347 B.c.)
Plato thought Reality was not to be found in the physical world (the
world of sensibilia) at all but rather in the nonphysical world of intel ligibilia , a world that housed mathematical entities and other nonphysical objects of thought . Only by turnin.g away from the sensible
world and by contemplating the nonphysical objects of intellection
could real understanding be achieved. Mathematics was offered as
the model of Knowledge and how it was to be won . Interaction with
the physical world might yield opinion or mere belief, but it could
never yield Knowledge of Reality . Pure contemplation was the proper
occupation of the rational soul, itself a nonphysical substance capable
of existence independent of the physical body . Since the theory postulates two distinct kinds of substances, mind -stuff and materialstuff , it is known as substancedualism.
Looked at in one way , Plato's theory is a desperate one. Faced with
the difficulty of specifying how to distinguish reliable observations
from illusory ones, good empirical theory from bad empirical theory ,
he turned his back on the muck and tangle that is the world and
dismissed it all as mere appearance and not worth the trouble . Exercise of reason alone was Plato' s avenue to knowledge . Although his
position was an extreme one, certain of its themes have survived with
surprising vitality . In particular , the idea that the mind has as its
objects of knowledge and contemplation nonphysical intelligibilia has
reappeared in countless guises as an argument for the special nature
of the mental , for dualism of various types, for the mind 's special
access to itself , and for the impossibility of reducing psychology to
neuroscience. (See chapters 7 and 8.)
Plato is the archetypal antinaturalist and contrasts vividly with contemporary naturalists who argue that the mind is the brain and that
empirical science is indispensable to discovering the nature of the
mind -brain , including the nature of the " Pure Reason" that Plato
thought would lead us to truth . The greatest erosion to Plato's views
was caused by the successof empirical science itself , and this erosion
began on a grand scale during the Renaissance.
243
With the resurgence of scientific progress after the Dark Ages , there
. The
soul
and
its contents
were
self - evident
to the soul ,
and thus the soul knew itself through self-examination - that is,
through introspection . Descartes believed that " I think , therefore I
am " (Cogito ergo sum ) could constitute the foundation of the edifice
of knowledge .
The distinction
the soul' s knowledge of itself ; in matters concerning the mind and its
contents , what seemed real to the mind was real . Having established
it conformed to an innately configured structure of ideas and princi ples , and it could distinguish
evident
. It
was , after
all , a rational
soul
with
of
innate
Plato , Descartes
was
a substance
dualist
, and
he took
the
soul
on the complexity
of
erate before action and to defer gratification , all suggested to Descartes that there was something entirely unique about the mind . So
remarkable
245
could have been false. For example, Pierre Trudeau is in fact bald,
but there is nothing inherently contradictory about the statement
" Trudeau
is not
bald . "
and for
analytic matters generally . In contrast to Descartes and the rational ists , they claimed that a priori reasoning cannot reveal truths about
the
nature
of
the
world
about
, existence
causation, the nature of the mind , and so forth . These matters, they
argued, are matters of fact and hence are matters for empirical
research
It was the battle cry of empiricism that only observations and experimental reasoning can lead to factual knowledge . For mathematical, geometrical, and logical know ledge, however , a priori reasoning
is appropriate , and confirmation
by experimentation
is irrelevant .2
havoc
must
we make
? If we take
in our
hand
any
volume
of
? No . Commit
it then
to the
flames
desire to characterize a method whereby respectable, worthy discoveries in science could be distinguished from mere speculation and idle
theorizing , and from II authoritative ' ! deliverances of sages and theo -
Introduction
and History
247
justification of our beliefs about the external world may be impos sible . Deductive justification is inappropriate , but inductive justifica tion cannot get started without observation of the causes, and yet
observation is always the effect of the causes we need to observe .
Second , Hume asked what became the tormenting question for em piricism : How is II experimental reasoning " about causes and effects
itself justified ? In terms of deduction ? That is impossible since the
conclusions of inductive arguments are not deductively derivable
from their premises . In terms of experimental reasoning ? That , Hume
pointed out , is arguing in a circle .
On one reading , Hume is an unqualified naturalist . He thought
that much of our representation /knowledge system is neither gov :"
erned nor organized by principles of logic , and that if we want to
understand
the cognitive
domain , we must look to other nonlogical - principles that will explain the causal factors involved .
For example , he thought that we are causedto suppose that our sense data faithfully represent reality ; it is simply one aspect - an empiri cally explorable aspect - of the way the human mind works . This is
interesting because it implies a criticism concerning the adequacy of a
science of the mind that assumes that cognition is fundamentally
logical (and language -like ) in nature . To understand how the mind
represents , and how representations are transformed and recon figured , it will be necessary to investigate scientifically the causal
principles that in fact govern the mind ' s operation . (A modern ver sion of such criticism of the mind -as-logic -machine idea will be con sidered in chapters 9 and 10.)
Some of the importance of Hume ' s contribution to the empiricist
theory of knowledge derives from his careful and ruthless working
out of the implications of empiricist principles and from his discovery
of profound and intractable problems that arise within the empiricist
framework . In coping with these problems , later thinkers were even tually forced to abandon much of that framework , including the idea
that science must have an absolute foundation of certainties given in '
sensory experience , that induction is the only principle available for
empirical justification , that there is an absolute distinction between
truths of language and truths of fact , and that the mind can know its
nature through introspection . Important in the progress towards
these revisions was Immanuel Kant .
Kant (1724- 1804)
If the problem is to show how the mind may contain a faithful copy of
Reality , as something knowable without mediation by the mind (or ,
as we would say , the brain ), then Kant , agreeing with Hume , said the
249
for
logic
erful
if
and
hence
influence
one
has
tions
as
to
cognitive
laws
theory
of
it
theme
occurred
be
are
raised
our
in
but
( 1839
Many
it
1914
notion
and
that
only
reality
me
more
Sanders
ery
he
idea
and
inquiry
will
science
for
tion
it
fact
is
Peirce
an
that
ence
not
believed
we
be
In
we
the
investigated
the
of
himself
.
the
we
distinction
and
Charles
.
there
is
we
the
reality
no
was
an
progressing
the
the
other
call
limit
of
completed
Truth
a
science
at
run
now
Peirce
tomfool
science
long
between
and
scientific
no
realist
idealized
concep
though
we
appearance
progressively
the
more
distinction
can
point
to
knower
ourselves
are
is
what
science
that
interests
may
not
metaphysical
want
which
pragmatism
reality
' s
the
metaphysical
idealized
what
Kant
and
America
was
is
in
espoused
construction
of
only
reality
was
of
end
by
avail
by
world
Kant
which
thread
nature
In
he
becoming
view
a
if
treat
must
his
from
that
must
which
that
one
.
,
task
inspired
richness
' s
in
the
about
that
mind
different
pioneer
that
truth
toward
reach
the
A
pragmatists
completability
argued
evolves
can
their
outer
work
Reality
description
the
evolving
the
nature
. Accordingly
ever
beliefs
improved
the
that
his
and
visionary
view
the
about
him
Unknowable
the
true
Reality
and
of
that
say
is
other
the
instead
discovers
be
view
idealism
but
reality
to
was
took
it
for
our
to
can
that
question
about
that
was
nought
mental
connected
the
If
adequate
complexity
thread
is
is
Peirce
thought
the
such
world
is
not
view
beliefs
was
a
from
one
the
the
This
epis
concepts
the
world
our
nevertheless
lead
inner
that
not
Feyerabend
then
are
the
trend
of
of
threads
thoughts
our
way
conforming
is
reconfigured
concepts
about
science
was
be
concepts
,
-
revolutionary
as
framework
might
pow
much
priori
a
( such
by
same
by
,
Peirce
to
the
as
that
provoke
entire
those
beliefs
much
improved
adopted
life
whether
whether
science
is
Evidently
representa
knowledge
to
mediated
mental
assume
processes
the
sentences
to
philosophers
inner
ourselves
always
its
that
our
about
and
on
tend
was
mind
the
operations
will
mind
of
introspective
later
the
characterizes
began
possibility
understanding
of
philosophy
one
that
to
the
are
mind
privileged
implies
and
the
then
the
Kantian
when
we
operations
that
and
logic
temologically
logical
- American
- like
of
investigation
The
the
British
sentence
the
in
for
on
knowing
means
of
all
whatever
and
that
in
in
science
within
its
sci
,
something
investigations
is
relevant
he
acquisition
as
such
is
science
. Moreover
community
and
reality
as
only
science
knowledge
the
scientific
drawn
outside
understand
and
be
and
accurate
250
251
l and History
and
therefore
dence
interpretations
are
supporting
and
them
James
of
had
views
certainties
In
any
case
brash
and
until
or
of
to
the
philosophers
by
the
the
not
new
on
These
take
new
time
of
seemed
in
philosophy
and
other
other
important
meaning
understand
their
Hanson
were
theories
to
revolu
of
hold
Feyerabend
bedrock
they
there
empiricists
rather
ahead
not
evi
progress
rests
perhaps
meantime
logic
logical
it
reason
did
Quine
the
empirical
makes
philosophers
ideas
In
the
because
of
as
science
Continental
Kuhn
science
developments
is
good
that
were
their
of
of
tempts
and
work
it
is
certainties
and
later
pragmatists
British
Yankee
the
as
it
that
sense
thoughts
only
However
about
of
tionary
what
and
the
made
at
science
.
SCIence
Frege
1848
Gottlob
this
logic
was
of
naturalist
also
see
ably
cally
It
the
important
rigorous
proved
from
logical
systems
provided
symbolic
sis
whole
Although
eventually
state
using
the
of
around
logic
mathematics
took
and
on
some
ominous
be
residual
proportions
reduced
to
the
all
the
crystal
clear
to
under
to
logic
demonstration
show
and
the
analy
deep
used
primitive
from
with
the
relations
Whitehead
just
problems
that
between
namely
later
in
say
transformed
and
relations
from
1944
sentences
to
vanished
Russell
be
of
of
it
were
were
can
structures
simply
provided
sentence
truth
propelled
spectacular
It
Tarski
the
logi
characteriz
construction
suffice
system
others
could
were
the
been
problems
truly
the
the
problems
something
there
in
manageable
do
where
for
theorems
details
of
and
as
of
hoary
new
to
to
and
logic
the
configuration
suddenly
well
how
of
system
Certain
as
incompar
syllogistic
permitted
into
the
grand
getting
rather
going
and
for
had
mysteries
logic
Without
means
Logic
rules
was
sentences
it
mind
sentences
characterization
formed
modern
old
symbolizing
and
the
collaborator
that
the
lending
the
of
his
logic
of
axioms
structures
the
of
characterization
mathematical
system
structures
and
for
sentences
in
and
than
sentences
of
theory
within
other
logical
new
in
discouraging
nature
Russell
between
mathematical
the
epistemology
in
struc
consisted
hence
Bertrand
relations
on
to
As
re
justificatory
effects
and
means
structures
the
those
illuminating
provided
the
and
and
important
ing
of
philosophers
effect
epistemology
devised
powerful
Aristotle
some
logic
how
depicting
profound
later
Whitehead
mathematical
on
theme
somewhat
more
of
effects
had
to
and
1971
modern
antinaturalist
approach
Frege
profound
shall
the
1872
program
it
we
to
have
As
support
empiricist
science
general
Russell
invented
to
the
ture
and
essentially
juvenated
1925
Frege
their
how
set
the
theory
reduction
that
that
one
253
. A
synthesis
in
namics
the
completed
theory
was
, and
ble
that
in
general
science
the nature
cation
might
the
consequence
and
the
logical
instances
relations
Logical
empiricism
owed
positivism
, which
losophers
of science
positivism
as a cohesive
convening
of the
became
known
1938 , and
erally
the
grown
was
the
weak
its
to
means
, and
science
suspicion
insight
describe
1970
newer
much
and
century
the
that
regular
eventually
dissolved
knowledge
logical
in
gen and
empiricism
progeny
had
changed
, what
over .
endured
theory
, and
in
had
taken
justified
, religion
phi -
. Logical
philosophy
scientifically
logical
and
to an end . Logical
within
its
to
formally
science
both
of superstition
, and
metaphysical
Empiricism
of the
the
of modem
logical
formal
logical
about
the
structure
system
ances
of the
tences
; the
and
observations
sensory
features
certitude
argued
for
set
that
science
that
of
the
and
concept
and
be possi beliefs
by
of " science
developed
a theory
in general . The
deliver
yielded
of observations
observation
it might
scientific
empiricists
of knowledge
were
was
the
. By widening
, the
senses
in
logic
all knowledge
sheerly
empiricists
systematicity
to include
foundational
laws
with
came
were
and
, between
society
about
influence
principles
for
. The
. By roughly
toothless
enthusiasm
correlative
in earnest
the
show
vitality
of the
could
in 1928 , a group
Circle
its
enormous
and
turn
justifi
inevitable
by scientists
as a movement
of theories
heir , and
evidence
original
began
society
positivism
central
maundering
ble
Mach
its
that
suggested
: one
superstructures
the
movement
Ernst
was
of
an
Russell
established
after
as the Vienna
its
sciences
much
in Vienna
and
about
and
was
possi -
. Reflection
of science
and
a movement
, as a skein
, was
Though
was
logical
empiricism
theories
foundations
to have
Helmholtz
so on , and
, methodology
and
revo -
. It seemed
explanations
out
there
. The
by
progress
Frege
quantum
possible
worked
of
the
old , and
, and
in an almost
physiology
excitement
system
between
, between
it seem
classical
thermody
for physics
, learning
and
logic
a logical
the
humans
scientific
, the
of making
and
vision
laws
and
replaced
in
the
physics
relatj .vity
made
work
how
of
of
, and
left
possibilities
enormous
explain
nature
. Third
possibility
, and
explain
of science
, and
the
generated
, it might
of
that
by Darwin
of biology
others
to have
theories
about
wrought
fragmentation
theory
new
excitement
science
the
, electrodynamics
seemed
special
powerful
in biology
genuine
was
that
state . The
renewed
Pavlov
factor
of mechanics
synthesis
were
lution
second
physics
observation
were
sentences
knowledge
generally
sen -
referred
provide
"
to
the
, and
Introduction
and History
255
257
observable
effects
were considered
meaningless
on this criterion ,
even though they might appear to the unwary to have meaning . The
logical empiricists ' point was that if a sentence had no empirical consequenceswhatever , if its being true or false could make no possible
difference to any experience one might have in the world , then the
sentence was without content , though by virtue of its grammatical
structure it might masquerade as meaningful . Sentences drawn from
they had a problem of their own : if they could not reduce the meaning of respectable theoretical expressions to basic observations , these
on
this
very
issue
Reductionwas the focus of concentrated effort , because logical empiricists typically had the view that science should constitute a
unified , coherent body of knowledge . Ultimately , they argued, less
basic sciencessuch as chemistry should reduce to more basic sciences
such as physics, and in the fullness of time an integrated account,
including sciencesof brain and behavior , ought to emerge. Within the
framework of the logical empiricist program , philosophers of science
tried to figure out a precise theory of what reduction is and what
conditions
must
be
satisfied
to effect
a reduction
. The
account
en -
Introduction
and History
259
Broca's area and no speech loss, and whether there is speech loss
with lesions elsewhere. Failure to falsify will then be significant , un like the collection of verifying cases. Popper's claim was that if the
scientist accepts hypotheses by finding confirming instances, he will
end up believing a great many false hypotheses and following a great
many dead ends. On the other hand , if he has a hypothesis that has
withstood tough attempts at falsification , then he can accept that
hypothesis - not as true , not as confirmed , but as the best hypothesis
available so far . Because he had a wholly different picture of
justification , Popper came to quite different views about the dynam ics and structure of science , and of knowledge
generally .
Furthermore , Popper disagreed with the assumption that what scientists should try to formulate are explanatory hypotheses with high
probability . On the contrary , he said, hypotheses are interesting only
if they are bold - that is, if they are improbable and thus likely to be
falsified . For then , to withstand
falsification
by rigorous testing is a
complex affair than it seems, and that contrary to the dream of logical
empiricism the falsification of a hypothesis is not a straightforward
affair , with logic our only counsel. Quine was inspired by the much
earlier work of Pierre Duhem (1861- 1916), who argued that hypoth eses cannot
be tested
in isolation
from
the
whole
theoretical
network
in which they figure and hence that the idea of " crucial experiments"
was far more messy than generally assumed. Quine was impressed
by the force of Duhem ' s arguments and saw that they had radical
implications
for the logical empiricist
theories of explanation ,
confirmation , and reduction . Duhem perceived that in practice , inter -
261
hypothesis H and auxiliary assumptions A are true, then 0 will happen. When in the circumstance 0 fails to happen, H is disconfirmed .
Duhem saw that the failure of 0 to happen could quite as well be
explained , not by H ' s falsity , but by the falsity of an auxiliary assumption . The experiment itself cannot tell us that it is H rather than one or
more of the auxiliary assumptions in A that is false. All that logic can
tell us
nothing
believe
a wider
given the failure of 0 , it might be some part of the background network that is to blame . As Duhem put it , we " can never subject an
Consider the following illustrations . Using Newton 's laws and auxiliary assumptions concerning the orbits, velocities, and massesof the
planets and the mass of the sun, astronomers predicted the orbit of
the newly discovered planet Uranus . Schematically, they said, " If N
and Al . . . An ' then 0 ." One of their auxiliary
assumptions
was
what they did observe about the orbit of Uranus , they predicted there
must be another planetary body close enough to Uranus to affect its
orbit . Shortly thereafter , the planet Neptune was discovered through
263
making suitable and sometimes large-scale adjustments in the background theoretical network . Any sentence can be safeguarded from
falsity and revision , so long as we are prepared to make appropriate
adjustments to other parts of the theoretical network . Conversely, no
sentence has special epistemic properties that protect it from revision
as neuronal
connections
network
as a whole
, and
where
the
revisions
should
occur
in the
in
our
theoretical
? How
do
we
know
where
to
assign blame? Evidently , in scientific practice a variety of considerations enter : if a hypothesis already explains a good deal and there is
no other theory with which to replace it , scientists are reluctant to see
it as the part to revise . If revising one deep feature in the theoretical
are
observations
and
rather
vague
observations
at that -
ples for making decisions about data, and will following them lead to
true theories? When should we be revolutionary in coping with data,
and when conservative? If we are consistently revolutionary , we will
never accumulate an " established " body of scientific knowledge ; yet
. Sometimes
tion judgment
it may
even
be reasonable
itself , or to reinterpret
to revise
an observa
, in some
sense
of reasonable
, since
in the
actual
business
of
running experiments, in the hurly -burly of the laboratory , observation judgments are never quite as secure and unimpeachable as they
are in philosophers ' theories of observation .
In sum, given the Quine -Duhem thesis, the difficulty then confronting philosophers of science was how to characterize methodol ogy in science- how to distinguish justified from unjustified
revisions , how to provide principles for rational conduct in scientific
inquiry . These were problems that clearly could not be solved within
the framework of logical empiricism . One view is that the answers are
to be found not by a priori pondering or by linguistic analysis but by
empirical investigation wherever relevant . If , by means of psychology , history of science , and neuroscience , we can determine how the
epistemic division
beliefs . As
Einstein
put
whole
of science
is
about external macro objects, about colors, about heat, about motion ,
and so on, are best seen as just some hypotheses among others that
the brain uses to theorize about the meager input
sensory periphery . Sometimes the theories humans collectively devise replace existing theories, presumably because they explain and
265
predict in superior fashion . Primitive theories give way to sophisticated theories, and as the latter become the common coin of
everyday life , they may then acquire the status of common sense. A
remarkable thing about the human brain is that it can use those primi tive theories to bootstrap its way to ever more comprehensive and
powerful theories- to find the reality behind the appearances. (For a
fuller discussion of this point , see chapters 7, 9, and 10.)
A consequence of Quine 's view is that even our epistemological
convictions about what it is to acquire knowledge and about the nature of explanation , justification , and confirmation - about the nature
of the scientific enterprise itself - are subject to revision and to correction . As we come to understand how the brain works , we will come
to learn about what it is for a brain to " theorize ," to be " rational ," and
to " understand ." We will discover general principles of brain operation that may change, and change radically , our existing epistemological conceptions .
It is in this sense that there is no first philosophy . There is no
corpus of philosophical doctrine concerning science and epistemology such that we can be sure it is the Truth to which all science must
conform . There is, as Quine remarks, no Archimedean point outside
all science from which we can pronounce upon the acceptability of
scientific theories . In abandoning the view that there is a First and
Inviolable Philosophy , Quine (1960:3) urges us to adopt a metaphor
coined by Neurath : " Science is like a boat, which we rebuild plank by
plank while staying afloat in it . The philosopher and the scientist are
in the same boat ."
What Happenedto the Verification Theoryof Meaning?
According to the verification theory of meaning, the meaning of a
nonobservation term was given by the set of empirical consequences
of sentences containing the term . From the outset, there were prob lems with this view . Suppose that a certain empirical consequence
depends on technology conceived but not yet invented . Is that part of
the meaning ? Suppose that a certain empirical consequence depends
on physics not yet established- or not yet conceived. Or suppose
that some boring little prediction is falsified . Does that change the
meaning of the term ? Suppose that sentences have empirical consequences that only occur to us tomorrow . Are they part of the meaning ? And so forth . Moreover , it emerged that the meaning of the most
respectable of theoretical terms was defined implicitly by the theory
the terms figured in , not by the empirical consequencesof the theory .
Terms such as " force field ," " energy," and " electromagnetic radiation " were prime examples where meaning was a function of the
266
267
we would do best to divest ourselves of (Quine 1953). Thus germinated the network theory of meaning . (See also chapters 8 and 9.)
What Happenedto the Empirical Foundations?
From the standpoint of logical empiricism the nature and epistemological role of observation sentences were unproblematic . Observation sentences were known through experience, and they were the
foundational " Given" of knowledge structures . To be known through
observation seemed pretty straightforward , and to be justified solely
on the basis of that observation seemed equally straightforward .
Theoretical sentences, on the other hand , were thought to be decidedly problematic , with respect both to meaning and to justification . However , the Quine -Ouhem thesis raised the possibility
that sometimes, in order to maximize coherence and explanatory
power , revisions of sentences at the observation level may be necessary. Sometimes, perhaps, the foundational basis should be
reconfigured if the advance in theory is sufficient to require it . Yet if
the foundational sentences can be revised, they cannot serve as the
enduring bedrock of science, on which theoretical superstructure is
supposedly built .
Seeing the implications of the Quine -Ouhem thesis, a number of
philosophers began to examine what had seemed unproblematic ,
namely, observation sentences. Particularly influential was the work
of Paul Feyerabend (1963a), N . R. Hanson (1967), and Mary Hesse
(1970). Hesse pointed out that for knowledge structures to have a
foundation of observation sentences that stand firm and inviolable
while the theoretical sentences above them come and go, observation
sentences must be independent of theoretical sentences. She therefore examined whether they were indeed independent , and hence
whether there could be an independent observation language.
To begin with , it turned out that logical empiricism had no very
clear definition of just what an observation sentence is. Sometimes it
was defined as a sentence known through observation, but such a
definition is blatantly circular . " Known directly through experience"
is just as bad. " An observation predicate is one directly applicable to
the empirical world " is fine except that it will not pick out predicates
whose application is free from theory .
The point is, many theoretical predicates can be directly noninferentially - applied to the world once the speaker's knowledge
of the embedding theory is second nature to him . An experienced
geologist can see that the term " glaciation" is applicable to a certain
geological feature; an experienced neurologist can see, without going
through intervening inferential steps, that a patient has Parkinson's
268
disease; a seasoned farmer can see fire blight on pear trees; an astronomer can see red shifts; a physicist can see alpha particles in a
bubble chamber; and so forth .
The other side of the coin is this : even if observation predicates are
initially learned by direct association with properties in the world ,
their meaning is not independent of laws regarding such properties .
This, argued Hesse, is a straightforward consequence of two things :
(1) The direct association of predicates with types of empirical
situations inevitably leaves a good deal of slack in what counts as
an appropriate instance of the relevant type of situation . Situations are always complex and various; similarity is not a transitive relation ; and the learning paradigms are often ill chosen.
(2) The application of any predicate must therefore also be conditioned by some rough generalizations or laws, which relate the
real occurrence of that feature to the occurrence of other features.
Suppose that a child hears the word " dog" in the vicinity of Rover
and associates the word with something or other in that situation possibly with all and only the dog himself , possibly with the-dog-in this-room or with doggy -at-dinner , or with doggishness, or with the
dog' s tailor the dog' s bark, or even with furry things . The word
J'dog" eventually comes to be more narrowly defined to fit standard
usage not simply in virtue of the initial learning but in virtue of the
additional information the child acquires about dogs through hearing
the .word in other contexts. Reclassification may take place for the
child as a result of imparted generalizations about dogs. To take a
crude example, if he hears J'Dogs like table scraps," he will abandon a
classification that applied JIdog" to the wagging furry part of the
animal .
Adults experience analogous events. A beginning neuroscientist 's
first observation through a microscope may produce puzzlement - it
may be difficult to know what is artifact and what is part of the cell.
(J'That's endoplasmic reticulum ??" ) Theory informs observation , and
after a short while it becomes hard not to see, say, the end bulb .
Notice that if I see something as an end bulb , then I imply that a
whole range of additional properties will obtain : that it is at the end of
an axon, that if tested it would be fo1.ind to contain synaptic vesicles,
that if we looked at it under an electron microscope we would see
synapses, and so on . To apply the descriptive term J'end bulb " to
what is seen in the microscope is not a sheerI naked observation : it
implies an indefinite number of generalizations applicable to the object. This cascade of indefinitely many implications is a general fea-
269
. Since
sometimes
that
reclassification
will
result
from
pressure from generalizations, albeit folksy generalizations, the predicate cannot be considered immune from theory and hence is not
independen t of theory .
The
second
consideration
also
en tails
that
reclassification
can
take
wider conditions : feet on a wintry day, the stomach after too many
green apples, the skin after a bum , excessively loud sounds, an electric shock, headaches, loss of something loved . Are the sensations
evoked on the different occasions exactly alike? Evidently not . (See
also Churchland
and Churchland
situation
in
reclassifications
of a term in situations
which
it
was
learned
. That is , observation
terms
to
there
be
are
are bound
bound
to
to be revised .
seem
that
there
is a residual
core
that
is attached
to the
subjective experience or quale. " Warm ," it may be argued , does have
some of its meaning fixed by generalizations one believes , but some of
its meaning is the subjective experience of warmth itself . All the rest
could
be swept
away
still know
what
270
But the network of laws is the primary and basic determiner of meaning . Here is his test case: Suppose a species of hominid evolved with
eyes sensitive only -to wavelengths in the infrared range. Hence, they
detect temperature by means of the visual system. Consequently ,
they see warm things as whitish -grey, hot things as white , cool things
as gray, cold things as black, and so on . In observations, where we
use " hot " they use " crin ." For example, they say that a boiling kettle
is crin and an ice cube is not at all crin . Now , does " crin" mean " hot ,"
or does it mean " white " ?
If we translate on the assumption that " crin " means " hot ," it turns
out that the hominids espouse the same scientific laws concerning
heat and temperature as we do, and they make most of the same
perceptual discriminations we do . That is, ice cubes are cold and
boiling kettles are hot . Suppose that for some reason we need to
cooperate with them and hence that we need a good and efficient
translation of their language. Then we will undoubtedly translate
their " crin " as our " hot" and not as our " white ." Why ? Becausewe
need the translation to communicate, and what we care about most is
exchange of information - about, for example, whether something
has the property that boils water , not whether someone's qualia are
just like mine when he detects the property that boils water .
We might conceivably prefer the qualia-based translation , but if we
do, then the beliefs we must then ascribe to the hominids , instead of
being isomorphic with our beliefs, are very odd . On such a translation , they think that ice cubes are black, that white objects sizzle and
turn gray when splashed with river water (we of course think that hot
objects sizzle and turn cool), that gray objects turn white in the heat
of the sun, and that white objects cause a painful burn ; indeed, they
turn out to believe all manner of rather bizarre things . When they
shout the warning " It is crin , it is crin !" we take them to mean " It is
white , it is white !" - and since we see no white objects of a threatening nature about, we are likely to get burned . To maximize translational efficiency , we will ignore as irrelevant the experiential
differences between them and us. And so will they , since their translation of our beliefs, if guided by considerations of qualia, will ascribe
systematically bizarre beliefs to us. When push comes to shove, laws
are more important than qualia in determining meaning .
Empirical foundations of science and knowledge generally are not
absolute and forever fixed; rather , they are foundations only relative
to a particular encompassing network . What counts as the observable
base will evolve and change as theory evolves and changes. Any
observation sentence may, under pressure from evolving theory , be
seen as false. When philosophers came to this realization , the tradi -
271
272 ThePhilosophy
of Science
by asking what is and is not conceivable or imaginable , by assaying
whether something comports with the existing meaning of the words
in question , and by amassiIlg fine details on how " ordinary people"
use the words or on " our concept" of the matter at hand . Empirical
theories as well as philosophical theories about mental states and
processesare sometimes criticized on the basis of whether or not they
mesh with existing " ordinary usage" - as analyzed by the philoso pher . For example, a standard criticism of the idea that mental states
(e.g., pains, beliefs, perceptions ) are in fact identical with brain states
is the argument that brain states presumably occur at a certain time
and a certain place in the brain , but we do not speak of mental states
(e.g., a desire for tomato soup) as being at a specific spatial position ,
let alone at several inches behind the eyes.
Or it is pointed out that we speak of a pain as searing, and of a
belief as true , but it is alleged that J'It makes no sense to say that a
brain state is searing, or true ." It is urged that it is meaningless to say
that the mental states are brain states because our concept of mental
states is radically different from our concept of brain states. And so
on. This style of argument , long typified by philosophy at Oxford , is
on the wane even there, but it is still advanced quite seriously, and a
forthright specimen of the genre can be found in a recent book on
reduction and the mind by McGinn (1982). For example, he says,
It is rather that mental concepts are intuitively such that no physical concepts could characterize the essential nature of the mental
property denoted . In other words , it seems that mental concepts
alreadycontain the essenceof mental phenomena and that physi cal concepts are necessarily unsuited to this role; . . . . (p . 19)
The method is suspect. First, whether a hypothesis makes sense to
someone will not be independent of his background beliefs and assumptions . So what makes no sense to a dualist may make perfectly
good sense to a physicalist . Nor are usage and intuition in general
anywhere near as uniform and doctrinaire as the method must assume. Feyerabend was perhaps the most severe critic of the method ,
and bypassing the fashionable disputes, he went straight to the heart
of its assumptions :
The more people differ in their fundamental ideas, the more
difficult will it be to uncover such regularities . Hence analysis of
usage will work best in a closed society that is firmly held together by a powerful myth as was the philosophy in the Oxford
of about 10 years ago. (1963a:192 (Morick ))
273
invisible
to the
naked
eye
caused
disease
, but
others
found
it
states
. For
those
who
have
a framework
within
which
it is
cide anything about the actual empirical nature of what those concepts are believed to apply to. Feyerabend described the inadequacy
of appealing to the common idiom as a source of truth in the follow .
lng
way :
to compare it with
in
its
favor
. Also
there
are
many
facts
which
are
inacces
idiom
is in -
A common philosophical tactic for criticizing neurobiological hy potheses concerning the identification of mental processeswith brain
processesused to be to brand these hypotheses as " incoherent " or to
cite them , pityingly , as having
tripped
as a mere conceptual
or inappropriate
in
the application of categories depends tremendously on one' s empir ical beliefs and one ' s theoretical imagination
(Rorty 1965).
The important thing for getting at the truth about brains is not
whether in customary usage ordinary humans-in -the-street do say
that
persons
remember
but
do not
say that
brains
remember
; rather
, it
275
might tell us something about what we do believe, but " the question
of the truth of the beliefs has not been touched " (1963b:144 (Borst )).
Like Kant , Feyerabend realized that our apprehension of our inner
world is not more basic, privileged , or immediate than our apprehension of our outer world . Our understanding of our inner life is
mediated by the concepts we use in apprehending mental states and
processes. But unlike Kant , Feyerabend thought the whole fabric of
mental concepts might be systematically improvable and revisable.
(Seealso Rorty 1965; Paul M . Churchland 1979, 1981.) The nature and
degree of doctrinal improvement is a matter for discovery in science,
not for protective and proprietary linguistic analysis. Even if our observation sentences describing mental states and processes are, relative to existing theory, foundational , they form no part of an absolute
foundation
The destruction of the logical empiricist theory of empirical founda tions and theory of meaning thus turned out to have profound impli cations for the philosophy of mind . It liberated philosophers both
from the constraints of holding fixed the current meaning of certain
words and from the limitations
philosophers
ical discoveries in psychology , neuroscience, artificial intelligence research, and so forth , could mold and shape and perhaps transmute
the language of the men tal .
Nonetheless, in an obvious sense and despite all the differences in
detail, the naturalistic bent of the new developments in the philoso phy of mind is akin not only to the pragmatism of Peirce and James
but also to the spirit of empiricism
the
cardinal
hunch
is that
to
discover
our
nature
we
must
see our -
and means. Hume 's enthusiasm for a science of man has a decidedly
contemporary
ring :
Chapter 7
Reduction and the Mind -Body Problem
observation.II
Paul K . Feyerabend, 1963
7.1 Introduction
In concluding the last chapter, I foreshadowed the implications for
philosophy of mind of the major revisions in the platform features of
logical empiricism . These changes permitted the development of a
naturalistic conception that envisioned research on the mind -brain as
an empiricalinvestigation of the nature of mental states and processes,
their causes, and their effects. Naturalism follows hard upon the
heels of the understanding that there is no first philosophy . Inevit ably the naturalistic approach leads us to inquire into the possibility
of a unified theory of the mind -brain , wherein psychological states
and processes are explained in terms of neuronal states and processes. A fundamental question concerning this possibility can be put
as follows : Can mental states and processesbe reducedto brain states
and processes? Can one be a reductionist ?
Not everyone expects mental states to reduce to brain states. On
the contrary , it has been my' observation that many philosophers and
cognitive scientists, most of the artificial intelligentsia , not a few
neuroscientists and biologists , and theologians generally, reject the
possibility as unlikely - and not merely as unlikely , but as flatly preposterous. But what , precisely, are they denying , and what is reduction ? The aim of this and the subsequent two chapters will be to
understand and evaluate the objections to reductionist strategies.
This is central to my program , for obviously , if reductionism is a
hopeless cause, then it would be foolish to search for an explanation
of mental states and processesin terms of brain states and processes.
if
reductionist
wondered
strategies
how
ogy
let
In
alone
to
chapter
who
consider
who
are
is
. 2
and
word
"
reduction
"
Reductionism
term
for
"
or
insult
"
"
"
they
right
is
"
in
begins
specifying
and
the
reduced
more
reduces
to
basic
"
Before
to
theory
amplify
and
the
prospects
it
will
is
that
be
wisest
in
have
agreed
acquire
would
be
con
matter
under
is
the
right
literature
is
have
"
debate
to
it
there
not
came
relevant
no
in
other
decided
here
certain
relation
word
to
take
my
by
that
are
reduction
is
simply
Statements
one
below
phenomenon
first
to
PR
upon
the
first
theory
specified
derivative
the
more
reduces
to
the
these
points
intertheoretic
reduction
focus
word
about
reduction
of
to
Most
characterizes
of
not
misunderstanding
theories
second
illustrate
for
diversity
specified
relevant
PB
nature
vague
communism
subject
be
bourgeois
inappropriate
the
and
"
perhaps
the
the
as
synonym
such
the
against
that
in
the
clarify
lay
stands
theory
"
"
sides
and
to
guard
phenomenon
the
in
Therefore
intend
"
With
"
for
science
to
"
avoid
that
the
they
ken
sense
between
that
science
used
as
reductionism
used
and
used
opposing
so
the
"
characterizes
the
what
in
be
because
in
reduction
relation
not
"
however
"
is
my
basic
that
assess
Here
applications
materialism
other
How
beyond
"
another
claim
theory
of
hounding
"
atheism
as
theory
another
often
appropriate
of
foremost
the
vivisectionism
"
of
meaning
sense
"
each
reduction
the
"
and
philosophy
be
"
"
traditionally
the
to
with
In
chapter
to
vague
as
and
is
word
the
sins
and
is
reduction
the
and
that
:se
contemporary
quarters
it
of
altogether
dilemma
It
case
about
word
The
consideration
by
disagree
the
notations
chances
In
account
of
some
Sometimes
diverse
uses
drop
science
an
range
indiscriminate
offering
in
go
what
word
state
movement
come
is
such
inevitable
reductionism
to
tho
distin
by
bewildering
abuse
sociobiology
is
such
by
experimentalism
"
equivocation
best
properties
physical
by
has
which
for
upon
synonym
militarism
and
"
has
"
of
capitalism
on
"
behaviorism
as
be
psychol
and
of
functionalist
not
well
in
formulated
the
is
may
Reduction
literature
general
of
chapters
it
it
dualists
set
sort
of
those
what
Intertheoretic
The
by
irreducible
any
part
for
reduction
order
from
then
research
arguments
groundwork
an
antireductionist
to
posed
be
states
philosophers
absurd
relevant
objections
to
mental
the
be
questions
there
consider
"
are
can
philosophical
consider
guishing
really
neuroscience
psychological
on
the
tactics
and
to
theories
very
case
to
at
issue
begin
to
neuro
In
namely
279
281
that exactly fits the logical empiricist pattern of reduction , and some
outstanding
(Hooker
1981 ) .
to get something
duced (as for example in the case of reducing the theory of optics to
electromagnetic theory ), but in other cases so much correction is
needed that almost nothing save a few low -level , homey " generaliza -
tions can be retained . Phlogiston theory of combustion and the demonic possession theory of nervous disorders are two such examples.
In these casesthe correction required was so thorough that it seemed
more appropriate to think of the old theoretical ontology as displaced
entirely by the new theoretical ontology . The spectrum accordingly
has at one end reduced theories that have been largely retained after
the reduction and at the other end theories that have been largely
rived from the new theory , then the logical empiricist conception of
reduction needs modification . In particular , the expectation that the
bridge principles are always necessary , and that their function is to
correlate phenomena in the old theory with phenomena in the new ,
282 ThePhilosophy
of Science
must be reconfigured . And this is a reconfiguring of enormous importance in thinking about reduction vis-a-vis psychology and
neuroscience. Should the old theory be largely displaced, then bridge
laws are typically dispensed with . We do not identify phlogiston with
oxygen, nor do we identify the demons that were said to possess
shaking palsy patients with dopamine deficiency . Presumably we
could do so, inasmuch as there is no strictly formal (logical) reason
not to, but there is no theoretical utility in these identifications . On
the contrary , such identifications might perpetuate misunderstand ing of what the new theories do . Instead, we say there is no such
thing as phlogiston , and there are no such things as demons. In these
cases, therefore , bridge principles play no role whatever .
In other caseswhere correlations are usefully forged, properties in
the new theory are matched up with properties in the correctedversion
of the old theory . However , the reconfigured properties of the corrected version of the old theory are never more than approximations
to the properties in the original theory . The rub is, therefore , that
whenever the corrections to the old theory are anything more than
relatively minor , it is always tendentious to claim that phenomena in
the old theory are to be identified with phenomena in the reducing
theory .
This dimension of scientific dynamics is important for our considerations becausemany standard objections to the possibility of a unified
theory of the mind -brain depend on the assumption that the commonsense understanding of mental states- of consciousness, beliefs,
desires, and so forth - is correct. Using this assumption , it is often
argued that no identification between neuronal states and mental
states is forthcoming . In a sense, this conviction too can be seen as a
legacy of logical empiricism , which cited immediate experience as
Given and hence as the certain foundation supporting the edifice of
knowledge . In seeing that reductions typically involve revision and
modification of the reduced theory , and that even outright elimina tion as well as reduction is a possibility , we should acknowledge the
necessity of relaxing our conviction concerning the basic correctness
of our understanding of the mental . To appreciate the kinds of possibilities inherent in the future evolution in neurobiological theory ,
we should consider how philosophers of science have improved
upon the logical empiricist story .
If the logical empiricist account of reduction is procrustean and
wrong , how then should the conditions for intertheoretic reduction
be specified? Roughly , as follo ,"Is : Within the new , reducing theory
T 8, construct an analogueT~ of the laws, etc., of the theory that is to be
reduced, TR' The analogue T~ can then be logically deduced from the
283
285
sort of co-evolution . As things stand, molecular genetics has not completely reduced transmission genetics, but its development toward a
richer theory that can effect the reduction is guided in part by results
and hypotheses in transmission genetics, whose research is in 'turn
inspired in part by questions arising from molecular genetics. In the
course of their co-evolution it has become evident that the " genes" as
characterized in early transmission genetics, likened unto beads on a
string , do not exist, and consequently the generalizations of transmission genetics have had to accommodate the molecular discoveries.
(See especially Philip Kitcher 1982; see also chapter 9.)
In chapter 9, where the co-evolution of theories will be discussed in
more detail , we will see that early signs of such mutual influencing
are now to be found in research on memory as neurobiologists , psychologists, and neurologists inform and correct one another . And in
chapter 10 I shall present an example of the co-evolution , albeit in the
early stages, of neurobiological and psychological hypotheses concerning the nature of attention . ~oth the reductive hypotheses from
neurobiology and the behavioral observations from psychology are
essential to the wider program of figuring out how the information
storage and retrieval system(s) works and how each corrects and
informs the other . It is of course conceivable that a richly developed
psychology of memory will never be reduced by a richly developed
neurobiology of memory , nor transmission genetics by molecular genetics, however the pairs co-evolve . On the other hand , it would be
fallacious to argue and reckless to predict that since the reduction is
not now available, it never will be.2 (The genetics example is further
discussed in chapter 9.)
It is often in this gradual co-evolutionary development of theories
that the corrections and extensions to both theories are made, and
from such theoretical interanimation may ultimately emerge a unified
theoretical framework . At this final stage the reduced theory will have
an explanation - of its properties , laws, entities, and so forth - in
terms of the reducing theory . However , as Francis Crick has remarked (in conversation ), by the time we get to the point of being
able to sit down and crank out the derivation of one theory! from
another, most of the really exciting science is over- the inspired
modeling , the wild and woolly imaginative flights , the wall -tumbling
experiments, the fitting and revising and revolutionizing are pretty
much behind us.
Admittedly , this is something of an exaggeration, inasmuch as
scientific theories are ever incomplete , and there is always fun to be
had. Nonetheless, this view is correct in emphasizing the importance
of the co-evolutionary process in achieving a reductively integrated
286 ThePhilosophy
of Science
theory . The logical empiricists , in focusing exclusively on the final
products of a long history of theoretical evolution , overlooked the
dynamics of theoretical evolution . This is a serious oversight , since it
is typically in a theory ' s evolution that the major reductive links are
forged and the major revisions - categorial and ontological - are
wrought .
Reductive achievements
mechanics
has succeeded
in
more complex atoms are emergent in some spooky sense, but only
that for want of mathematical solutions , an approximation will have
to suffice . In any case, whether the appropriate mathematics will be
, acceleration
, and
so on .
The appearance of retention , and thus the appearance of identify ing phenomena in one theory with phenomena in the other , is deceptive , however
, and
the
differences
between
the
two
theories
are
287
CM
Speaking more strictly , the deduction explains why CM was as successful as it was, within the classical limits . And the deduction does
show that SIR has the resources to " take over" the explanatory and
predictive jobs performed in the past by the original CM . That is why
this case still counts as a successful intertheoretic reduction , despite
the lack of a systematic set of intertheoretic identity statements.
A concern may yet linger : " What about mass itself - what about the
phenomenon that exists in nature ? Perhaps massCMis not identical to
massSTR
, but does mass itself reduce to massSTRor not ?" In order to
understand the question , we must know what the speaker means by
" mass." Suppose he says, " Well , it is that property which things
differentially have such that one thing is harder to push than another ,
288
such that some things weigh more than others, etc." Notice that in
giving the answer, in effect he gives a theory of sorts- a dinky little
theory perhaps, but his beliefs about what mass is are inevitably tied
to generalizations
world
, and
absolute
foundations
have
gone
to the wall
with
nature
, says
SIR , is relativistic
mass -
mass
as de -
What gets reduced are theories, and the stuff in the universe keeps
doing whatever it is doing while we theorize and theories come and
go . So the question " Does mass itself get reduced ?" does not make
any
sense .
pertinent to the mind -brain issue for the following reason: it is common for objections to the possibility of a theoretical unification of
neuroscience and psychology to lean heavily on our current intuitions
about our mental
of concepts
, " observational
" framework
. If it is irreducible
, that
may
be
" observable" concepts may be reconfigured or eliminated , and occupying the station of " intuitive framework " or " folk theory " pro-
289
290
Additionally , the distinction fundamental to impetus theory between being in motion and being at rest- is irrelevant to Newto nian dynamics , which instead draws a fundamental distinction
between moving at constant velocity (either at rest or at unchanging
speed) and moving at changing velocity (acclerating and decelerating ). In addition , the failure to see the differences between motion in
straight lines and motion deviating from straight lines was disastrous.
The corollary concerning curved impetus was a muck-up and can
easily be observed to be false- so easily, in fact, that the untroubled
claims to have observed verifying instances look like first -rate casesof
seeing what we believe .
The example is interesting for a second reason- namely , that
something rather like impetus theory seemsto be what many humans
uninformed of classical mechanics tend to believe and act upon , even
in the twentieth century . In a series of studies performed by McCloskey (1983), it was found that both in their behavior with regard to
moving objects and in their verbal predictions and explanations, students were far more Aristotelian than Newtonian . In one test college
subjects were asked to jog along and try to " bomb" a stationary target
by dropping a golf ball on it . Typically , they thought the ball would
fall straight down when released, and some even thought it would
move backward . Therefore, they did not drop the ball until the hand
holding it was directly over the target or even past it . In other experiments subjects tried to make a puck slide in a curved path by moving
their arm in a curved path before releasing the puck . Additionally ,
they explained projectile motion in terms of a force imparted to objects that keeps them in motion but slowly dissipates, after which
gravity takes over- an explanation reminiscent not of Newton but of
the medieval physicists . Among students acquainted with classical
mechanics there was a tendency to try to assimilate that theory to the
impetus theory , for example by construing momentum as a cause of
motion and in effect using " momentum " in much the same role in
which the old theory used " impetus ."
These experiments indicate that apart from the scientific community , most humans seem to employ a sort of folk physicsor, as McCloskey calls it , intuitive physics. Although a person might not describe his
set of beliefs concerning motion as a theory , given the role of such
beliefs in explaining and predicting motion , it is nonetheless appropriate to credit the person with a folk theory . The set of beliefs is not
likely the product of self-conscious theorizing , but insofar as they are
usedto explain and predict motion of objects, theyfunction for the personthe
lvaya consciouslyconstructedtheoryfunctions. Perhaps the folk theory is
casually acquired as one learns one's native language, or perhaps
Reduction
291
there is some innate structure in the brain that fastens on this theory
first . Whatever its origins , folk theory there does seem to be, and to
someone uninformed of Newton ' s laws the folk theory seems to be
self-evident and observably true - a bit of rock-solid commonsense.
Moreover , like earlier impetus theorists, ordinary folk think they
have observed what we know to be impossible trajectories for projectiles. (See McCloskey , Caramazza, and Green 1980, Clement 1982.)
To someone unacquainted with Newtonian theory the central wis dom of Aristotle ' s physics does seem intuitively dead obvious , and
Newton ' s first law may well seem counterintuitive and outlandish .
Only when it is experimentally shown how impetus theory and New ton' s laws predict different effects- only when it is demonstrated
how Newton 's laws can explain why impetus theory worked as well
as it did and how much more Newton 's laws could explain - does the
superiority of Newton 's laws begin to dislodge the old dogma root
and branch . Once the new theory becomes second nature, observations are made within its framework , not within the framework of the
abandoned theory . Unless the old theory is deliberately dislodged
and its falsity explicitly recognized, the inclination seems to be to
force the new theory to knit itself , somehow or other , in to the old
theory .
This example illustrates several points that will be useful in the
discussion of reducing mental states to brain states: first , that in reductive developments one theory can displace and falsify another;
second, that sometimes what is displaced and falsified is a folk theory
within which those who hold it make their observations; third , that
despite the self-evidence of the folk theory , it can be demonstrated to
be misconceived; and fourth , that as a newly acquired theory becomes familiar , it can be as routinely and casually used as the old folk
theory . For example, a seasoned engineer uses Newtonian mechanics
as effortlessly and " intuitively " as McCloskey' s subjects use impetus
theory . With routine use of a theory , one naturally makes economies
in the sequence of inferential steps and comes to be able to predict the
trajectory of a projectile without laborious deductions . One just
knows , and knows without conscious effort at Newtonian reasoning,
that a puck will not move in a curved path after release from a hand
swung in a circle, and it will seem faintly bizarre that anyone should
have thought otherwise .
A further pertinent observation is that even after a theory has been
displaced by a competing theory , some people may continue to
understand phenomena in terms of the displaced theory , quite as
though it were not false. Unless explicitly taught to think of motion in
a Newtonian framework , people seem disposed to employ the im -
293
it move
. Moreover
, one
cannot
observe
it or its
movements
though the apparent movement of the stars can be explained in New tonian theory by reference to the Earth's motion . In addition , of
course , the seemingly self -evident datum that the Earth is stationary
fate :
Harvey did not try to explain how animal spirits are concocted in the
heart
; instead
, he said
that
the
heart
is a pump
for blood
and
tried
to
mind
-brain
is this : what
we
take
to be obvious
or observational
within the framework of intuitive psychology is not guaranteed cor rectness or survivability simply on the strength of that obviousness
stakes are lower can we begin to get enough distance to see that our
" intuitive " conceptual framework for understanding ourselves, for
introspecting , for observing , for explaining and predicting behavior
may require revision or even elimination .
For the sake of completeness I should
of
294
scientific explanation I adopt here is a modified form of the deductive nomological (ON ) model worked out by the logical empiricists (chap ter 6; see also E. Nagel 1961) . The important modifications
are
twofold . (1) Intertheoretic reduction is an instance of ON explanation ,
where what is explained is not a single event but a law or set of laws .
Insofar as the account of intertheoretic reduction deviates from the
logical empiricist conception , there are corresponding modifications
in the theory of explanation . In particular , the theory must be
modified to allow for explananda to be reconfigured rather than deduced as they stand . The model accommodates this change without
strain . As noted earlier in the discussion of intertheoretic reduction ,
sometimes what is deduced is not the old law itself but a reconfigured
version or analogue of the old law . Accordingly , it is not the old law
that is explained , but its analogue . (2) Not all laws are universal
generalizations ; some are statistical generalizations . Thus , the basic
model must be broadened to encompass explanations where the cov ering laws are probabilistic . This modification will not in fact playa
significant role in the issues raised in this book , but to see how it
works , interested readers should consult Salmon 1971. (A great many
details concerning the deductive -nomological model of explanation
that I must ignore here are cogently discussed in E. Nagel 1961 and
Rescher 1970. For a more recent paper that answers many of the
objections to this model , see Cupples 1977.)
Having declared my preference for the deductive -nomological
theory of explanation , I emphatically add that I neither wish nor need
to be doctrinaire about this . On the contrary , I adopt this theor )' only
as a first approximation that enables us to discuss without folderol
the important aspects of intertheoretic reduction . Accordingly , the
framework for reduction and explanation is intended to accommo date explanations of the form described by Cummins (1983), in which
a capacity at one level of organization is explained in terms of the
coordinated functioning of capacities at a lower level of organization .
The points that I find central are that reduction involves explanation
and that typically reductions involve corrections of those levels of description that supply the explananda .
In any case I think the theory of explanation is bound to be provi sional , since I expect that as theories of brain function mature , we will
acquire a far deeper sense of what the brain is doing when it seeks
explanations , what it counts as explanatory fit , what understanding
comes to , and so on . And my conviction is that this deeper sense will
take us beyond the paradigm in which explanations in general are
characterized on the model of logical relation between sentences. (See
chapters 9 and 10.) Rather , it may hypothesize very different sorts of
295
from sentences .
are candidates
for
the
relevant
intertheoretic
reduction
. This
of reduction
(see for
example Sperry 1980, MacKay 1978, Eccles 1977), with the consequence that it is often hard to see what the issues are and what might
count
as evidence
one
way '" or
another
. In
these
circumstances
the
clash be-
important
generalizations
prehensive Theory - that delineates the psychological states and processesmediating perception , learning and memory , problem solving ,
cognitive mapping
is therefore
a great
and sensorimotor
capacities of organisms .
deal of research
between psychological
Granting the unfinished status of scientific psychology , it is nevertheless clear that for many of the generalizations
296
the
conditioned
stimulus
was
correlated
with
the
uncon
will
stimulus
L .
1979, McCormick
1984) .
Even so, some misgivings may linger about the possibility of reduction
should
it be assumed
that a reductive
strategy
means
an exclu -
for reduction .
involves
297
too has a place in the larger scheme of things . Only if one theory is
eliminated by another does it fall by the wayside .
Nor , it must be emphasized, does intertheoretic reduction imply an
essentially " compositional " research strategy . By a " compositional
strategy" is usually meant that one must first know all about the
micro properties
, then
what
is
the
issue ? Is
there
a residual
source
of
objections to intertheoretic reduction even after the story of the nature of intertheoretic reduction dispels some of the antireductionist
bugbears ? The answer is clearly yes . The residual problems are quite
complex, and they condense around two substantial points . Simplified , the first can be stated thus : " Our ordinary understanding of
our mental states provides the observational base against which the
and
."
The
of
the
reasons
supporting
commonsense
chapter 8.
The second condensation
for
this
conclusion
this general
framework
will
be
very
di -
sense of the
the
focus
of
and the
information
information
processing
involves
operations
on representa -
tions , and I agree that it must , then the quarrel is a singularly impor tant one, for it concerns the future development of that domain of
scientific psychology which addresses the cognitive capacities of
humans and other organisms . Moreover , this antireductionist
argu ment is relevant not only to the wider question of neuroscience -
against the unified conception are right , for the issue bears upon how
we decide to do research. If we are persuaded that a unified theory is
impossible and that some explanations of cognition , perception , and
so forth , cannot ever be explained in terms of neurobiological theory ,
then this profoundly affects how we conceive of our long -term research goals, and that affects our more immediate research strategies.
Which
research
research
visions
intuitions
to
abandon
to weed
and
out
which
and
to
which
nurture
to culture
, are
, which
affected
in
299
framework , its
. The
theme
of the
co - evolution
of scientific
theories
will
be the
I have
standing
referred
to
our
" commonsense
framework
for
under
about what is meant . For brevity ' s sake, I shall begin by replacing that
long-winded description with a shorter label, namely " folk psychology ~" Now by folk psychology I mean that rough -hewn set of
concepts , generalizations , and rules of thumb we all standardly use in
explaining and predicting human behavior . Folk psychology is commonsense psychology - the psychological lore in virtue of which we
explain behavior as the outcome of beliefs, desires, perceptions, expectations, goals, sensations, and so forth . It is a theory whose
generalizations
the
characterization
of the
mental
states
and
processes
referred
Reduction
301
nest Hemingway , though presumably they shared many general beliefs about why humans behave as they do. One way to begin to limn
the generalizations that figure in t,hecommonsense theory of human
behavior is to press for expansions of action explanations in order to
force the implicit and " understood " background assumptions into
the open . This is similar to the method used to limn folk physics. For
example, many people thought that Richard Nixon knew about the
cover-up of the Watergate break-in . Why ? Because, it will be said, he
was in the room when it was discussed. Why is that relevant ?
Crudely , because if someone x has normal hearing, and someone else
y says " p" in a normal speaking voice in x's vicinity , and x knows the
language, then x hears that p. Similarly , we may conclude that a
person was lying because he must have seen the body (anyone with
normal vision , in broad daylight , standing two feet away would
have), that someone did believe that there was money in the safe
(why else did he break in and jimmy the lock), and so forth . Fodor
(1981) suggests other instances: " seeing that a is F is a normal cause of
believing that a is F; statements that p are normally caused by beliefs
that p; the belief that a thing is red is a normal cause of the inference
that the thing is colored .. ." (p . 25) (For more discussion on the background generalizations , see Paul M . Churchland 1970, 1979.)
Those who find these generalizations to have an excessively cracker-barrel flavor should recall that they are up for discussion precisely
so that antireductionist arguments citing the essential correctness of
folk psychology can be addressed. Admittedly , the generalizations do
have a cracker-barrel quality , but this is characteristic of folk theory
generally, and it does not in the least imply that they are simpleminded . On the contrary , however folkish it is, folk psychology as a
framework of understanding is very complex. Moreover , folk psychology is where scientific psychology began, and if scientific psychology is to revise and improve upon folk psychology , it must know
what that framework consists of and whether and to what extent it
can be revised .
The insight that our psychological concepts are nodes in a background theory and that our explanations of human action proceed
within the framework of that theory is owed mainly to Sellars (1963)
and Feyerabend (1963b,c). It is an insight that has profoundly in fluenced research in the philosophy of mind . Together with the related work in epistemology and philosophy of science, the " theor)T"
theory of our commonsense beliefs about the mind has engendered
new understanding of the nature of mental concepts and rendered
tractable some baffling traditional questions, some of which pertain to
reduction .3 Before discussing how these insights can help in confront -
(" Pupil size of the person one is looking at affects beliefs about how
friendly the person is" ). The etiology of an individual 's possession of
a folk psychology is probably much like the etiology of his possession
thereby they come to attribute thoughts to each other . But this story
was designed not so much to be anthropologically reasonable as to
emphasize a logical point . Sellars ' s thesis was that beliefs and desires
are beliefsthat p and desiresthat p, and hence that the logic of such
expressions is importantly related to the logic of overt utterances
of " p . , ,4
303
rest as theories, what is important is not that they originated in selfconscious construction , but that in their explanatory and predictive
role theyfunction as theories. In this respect, folk psychology and folk
physics deserve to be called theories . Although it would be satisfying
to know the origins of commonsense lore and to know how any
individual comes to have the lore he has , our ignorance on these
matters in no way impugns the claim that the lore should be consid -
ered a theory in the defined sense. For that claim concerns the epistemological status and function of the lore, not its origins .
Folk Theoryand Rationality
Propositional attitudes are those mental states whose identity depends on a proposition specifying the content . If Jones has the belief
that hawks eat mice , then the proposition
the content or " object" of his belief- in other words , it specifies what
his belief is. Additionally , Jones might beafraid that hawks eat mice or
hope that hawks eat mice , or expect or wish or wonder whether hawks eat
mice , and in each case it is the proposition " Hawks eat mice " that
defines the content of the propositional attitude . (Pains , tickles , and
that obtain
hawks
belief
that
eat
mice
and
Leonard
eats
between
his
the content
belief
that
mice . And
Leonard
is a hawk
entail
his
so on .
and
desires
that
make
rational
sense
of the
behavior
. For
ex -
ample, Smith might say his reason for butchering the calves was that
the cost of feed was increasing and that the price of beef cattle was
declining . Smith ' s private deliberations must have gone something
like this : " The cost of feed is increasing , the beef cattle market is
my
income
, therefore
I should
butcher
' s last
theorem
, " and
the
decision
the
calves
no \,\T. " Re -
to butcher
the
calves
is no
contents
of the
mental
states
and
the
content
of the
decision
, and
some philosophers have wished to argue that this distinguishes absolutely explanations of intentional action from explanations in physics,
biology , and other branches of science. According to this view , even if
psychological
explanations
conform
304
pattern , it is not the deductive relation obtaining between the covering laws, initial conditions , and the explanandum that is important ;
rather, it is the internal , rational -in -the-light -of relations between the
contents of mental states that do the explanatory work .
In particular , this special relation has been cited as grounds for
rejecting the claim that explanations within folk psychology are
causal explanations and that the generalizations of folk psychology
are causal-explanatory . On the contrary , it has been argued, the
generalizations are not causal, and the style of explanation in which
they have a role is uniquely rational as opposed to causal. (See for
example Dray 1963.) What produces human behavior , it is said,
are
reasons
, not
causes
, and
we
understand
behavior
in
virtue
of
this
has
arguments
sometimes
functioned
as the
basis
for
antireductionist
The observation that in explanations of human behavior a rational ity relation connects statements about beliefs and desires with state -
that
beliefs
and
desires
therefore
do
not
cause
behavior
when
certain
causal
relations
obtain
between
machine
states ,
existence
of abstract
or formal
relations
defined
over
contents
Propositional Attitudes
x . . . has a maSSkg of n
. . . believes .that p
. . . has a chargecoulof n
. . . desires that p
. . . has a temperatureK
. . . perceives that p
of n
. . . has energYj of n
etc.
305
. . . hopes that p
etc.
Reduction
307
308
ThePhilosophyof Science
err, notice that it does not follow that no theory (no categorial
framework ) is employed . The remarkable accuracy could be owed to a
remarkably accurate application of a true theory in inner perception .
But the traditional
one could not be wrong - not merely that one mercifully , but contin gently , escaped error . If we could not be wrong , that allegedly would
be evidence for immediacy and directness.
The
trouble
is that
we
can
be
wrong
and
often
are . Sometimes
and desires
within
the confines
of a categorial
framework , and even though it is agreed that we can err in the appli cation of the categories of that framework , nevertheless
it may be
without
it may
that
we
cannot
be mistaken
that
we
are aware
world
, Descartes
found
a case
that
he
took
to
be
immune
to
I think
that
I am
aware
." Thus
Descartes
' s memorable
conclu
sion : " I think , therefore I am " (Cogito ergo sum ). Now the issue in
this
section
oneself
and
concerns
others
what is important
whether
within
one
understands
a theoretical
the
framework
mental
. For
states
this
of
issue ,
309
310 ThePhilosophy
of Science
ible with the " theory " theory of our commonsense understanding of
mental life . On the contrary , one's apprehension of one's own mental
states is a cognitive process that at some level exploits a conceptual
framework . The nature of that process, the degrees of plasticity it
may have, and the nature of the representations and their role in that
process remain to be empirically discovered by cognitive neurobiol ogy . Moreover , the notion of a conceptual or theoretical framework is
still only roughly delimited and will itself need to be revised, extended, or perhaps eventually displaced, by cognitive neurobiology . I
view the " theory " theory as a provisional conception that enables us
to disentangle ourse.lves from a confining epistemology and a
confining philosophy of mind . As neuroscience and scientific psychology collectively come to understand more about the nature of
information processing, so we shall come to understand what
theorizing is, and our inchoate conceptions will mature . Ultimately ,
to invoke again Neurath ' s metaphor , this plank in our boat will itself
be changed. This is not a philosophical flaw in the " theory " theory ; it
is just a frank acknowledgment that we do not understand much at all
about what it is for a brain to theorize .
Folk Psychologyand a NonphysicalMind
Finally , it may be argued against the " theory " theory of mental predi cates that if mental states are states of a nonphysical mind , then our
accessto them must be direct rather than theory -mediated . The argument is simply a non sequitur . Even if the mind should be nonphys ical, its accessto its states and processes may be via a theory , and it
may have lots of nonconscious mental reasoning and processing and
lots of nonconscious beliefs and desires. Moreover , even the nonphy sical mind may make mistakes and have false theories, one of which
may be about its own states and how they interact . Nonphys icalness in and of itself implies nothing about immediacy , directness,
and privileged access. The ontological question about what sort of
substance the mind is must be distinguished from the epistemological
question concerning how the mind knows itself .
7.4 Conclusions
lntertheoretic reduction is fundamentally a relation between theories,
such that one theory T 1 is said to reduce to another theory T2 if T 1 (or
an analogue of it ) is deducible from T2. One phenomenon PI is said to
reduce to another P2 if the theory that characterizes PI reduces to a
theory that characterizes P2. Reductions in science are only rarely
smooth reductions with uncomplicated cross-theoretical identifica -
311
tions . More typically , they are bumpy and thus involve varying
degrees of revision to the reduced science. Sometimes the correction
required is so massive that the candidate theory is better described
as having been displaced outright .
Accordingly , neuroscientists should not at this stage of their in quiries feel compelled to envision the details of a future reduction of
folk psychological generalizations , nor should they suppose themselves obliged to seek identifications between neurobiological states
and psychological states as characterized uJithin folk psychology
.
Identifications may be forthcoming , but then again they may not .
Whether they are is an empirical matter, not an a priori matter . The
crucial point is that identities are not required for reductions at a
remove from the retentive end of the spectrum . So unless neuroscientists can tell in advance that they will not find any reason either to
revise or to displace folk psychology , they are not honor bound to
design their research to turn up such identities .
The objections to the reductionist program that envisions a unified
theory of the mind -brain , in which psychological theory is ultimately
explained by neurobiological theory , derive from two sources: (1)
from a conviction that a subset of generalizations in scientific psychology are of such a nature as to resist reduction , and (2) from a conviction that folk psychology is both substantially correct and essentially
irreducible to neurobiology . The generalizations in scientific psychology that are thought to be problematic are typically in the domains of
cognitive psychology and social psychology and make reference to
beliefs, inferences, or, in general, propositional representations . Insofar as these problematic generalizations rely on a theory of representations drawn from folk psychology , these two patterns of
objection have a common element .
Although folk psychology has a profound familiarity and obviousness, and although the categories of folk psychology are observationally applied , it nevertheless remains true that folk psychology
is a theoretical framework and hence a framework whose adequacy
can be questioned and assessed. The adequacy of folk psychology is
in no way secured by its seeming to be overwhelmingly obvious., by
its being observationally applied , by its being applied introspectively ,
or even by its being innate , if such be the case. If we see that folk
psychology has no right to epistemological privilege , and no immu nity to revision and correction , then we can begin to see that its
generalizations and categories can be corrected and improved upon .
Indeed , an underlying theme in scientific psychology is that they can
and are being modified and even superseded.
312
of psychology
also not be enough for these arguments simply to point to a category ' s being observationally applied or seeming to be obvious . So much
is entirely compatible with the inadequacy and falsity of a theory .
Nor , of course, can the arguments quietly help themselves to as-'
sum ptions about the correctness of the generalizations and categories. The focus of chapter 8 will be a set of arguments pertaining to the
nature of our subjective experiences, and chapter 9 will examine and
evaluate the arguments based on the theory of the nature of
represen
ta tions
Chapter
Are
Mental
States
We
F . H
States
Irreducible
to
Neurobiological
are
deceived
. C . Crick
at
,
every
level
by
our
introspection
1979
8.1 Introduction
There are researchers in psychology , neuroscience, philosophy , and
other fields who have concluded that no unified theory of how the
mind -brain works will ever be devised and that at least some psychological phenomena are forever beyond the reductive reach of neuroscience. And some of these opinions survive the clarifying effect of
the account of reduction presented in chapter 7. Study of brains is all
very well , it may be allowed , but even in the long run it Vv
"ill not
explain much of significance concerning how we learn, remember,
speak, solve problems , and so on . Subjective experience, consciousness, reasoning, and even visual illusions are sometimes argued to be
forever beyond the explanatory reach of neuroscience, regardless of
unforeseen breakthroughs and surprising discoveries. No single reason rallies those who reject the possibility of a reduction , and there is
little unity among the skeptics concerning why a reduction is impossible. Rather, there is a spectrum of arguments , some of which are
deeply perplexing , others of which are less arguments than loving
expressions of which intuitions one has vowed to shelter, come what
may .
In this chapter and the next I want to examine those arguments that
seem especially persuasive or challenging , in order to determine
whether the skeptics are right . At the outset it is necessary to distin guish two very general patterns of argument . Arguments in the first
class are advanced by those I refer to as " boggled skeptics," who say
essentially that the brain is so horrendously complicated- there are
too many neurons and too many connections- that the hope of
understanding it is but a pipe dream. In all likelihood , the boggled
skeptic laments, the human brain is more complicated than it is
and
significance. Those who advance arguments of this type I call " prin cipled skeptics." (See also Patricia S. Churchland 1980a.)
The boggled skeptic may, of course, be right in his prediction .
Notice though that what he advances is a prediction , and one that
concerns the nature of the empirical world and whether , as a matter
of empirical fact, our brains are smart enough to invent an adequate
theory to explain certain empirical phenomena . Whether the boggled
skeptic is correct is an empirical matter , and so far at least there is no
empirical evidence to justify his gloomy prediction . So far as we can
tell now , the human brain might be more smart than it is complicated .
That neuroscience is hard does not imply that it is impossible , and
concerning
or done . Prob -
Improved technical means for investigating the brain have made possible
in
the
last
two
decades
research
and
results
that
would
have
directly discuss the case of the boggled skeptic, for the book as a
whole
constitutes
simis tic forecast
an indirect
invitation
to him
to reconsider
his
pes -
The arguments that I class together as defending principled skepticism are very diverse indeed , but they fall into roughly two antagonistic - groups . The first orients itself around
there
is a distinctive
mental
dimension
that
is not
reducible
to any -
317
319
Can the mind be affected by, say, electrical or magnetic fields ? For
Descartes, apparently not , for then it would have properties in common with matter , and its status as a radically different substance
would be imperiled .
On the classical picture , essentially two types of items were exchanged at the station where mind and brain interacted , wherever it
was, and these were sensations and volitions . The brain was thought
to send sensations to the mind , which could then use them in perception . The mind , on the other hand , was thought to send volitions to
the brain , which could then translate the volitions into motor effects.
The higher functions of the mind , including reasoning, consciousness, moral feelings, and the emotions , were assumed to function
independently of the brain , save for the extent to which perceptions
might figure in these functions . Perceptions were excepted because
they were to some extent dependent on sensations. This indepen dence of the higher mental operations from the physical business of
the brain was really the raison d'etre of the substance dualist hy pothesis, for it was these mental functions that seemed utterly inexplicable in material terms . Given a life of their own in the nonphysical
mind , reasoning and consciousness and their kind should be amenable to nonmaterial explanations, and getting these seemed far
easier than getting brain -based explanations . We shall see this theme
concerning reasoning and consciousness reappear in assorted guises
in virtually every antireductionist argument , including those most
recently minted .
The hypothesized independence of reasoning and consciousness
that makes substance dualism attractive is at the same time a chronic
and aggravating problem that costs it credibility . The difficulty is
straightforward : reasoning, consciousness, moral feelings, religious
feelings, political convictions , aesthetic judgments , moods, even
one's deep-seated personality traits - all can be affected if the brain is
affected by drugs or by lesions, for example. The more we know
about neurology and about neuropharmacology , the more evident it
is that the functions in question are not remotely as independent as
the classical hypothesis asserts. On the materialist hypothesis , the
observed interdependence is precisely what would be expected, but it
is distinctly embarrassing to the dualist hypothesis .
Recent hypotheses meant to explain the nature of the interaction
between the nonphysical mind and the physical brain are not
significant improvements upon Descartes's proposal . Although Eccles (1977, Eccles and Robinson 1984) has energetically addressed the
problem , his theory of the interaction remains metaphorical . His explanatory flow diagram consists essentially of many arrows connect-
320 ThePhilosophy
of Science
ing the " mind " box to the box for the language areas of the human
brain . The question that persists after study of the array of arrows is
this : what is the manner of interaction , and how does the nonphysical
mind bring about changes of state in the brain , and vice versa? The
inescapable conclusion is that the arrow -array is after all as much an
explanatory surd as the notion of Descartes's animal spirits finely but
mysteriously " affecting " the nonmaterial substance in the confines of
the pineal gland .
The unavailability of a solution to the manner of interaction between two radically different substances does not entail that substance
dualism
is false . For
all
we
know
now
, further
research
may
yet discover a solution . But with no leads at all and not even any
serious plans for finding a solution , it does mean that the hypothesis
has diminished
that the
not
exist
to be solved
alone
have
minds
, where
did
these
substances
come
from
?A
and unmotivated
excep -
tions to the plausible and unified story of the development of intelli gence provided by modern evolutionary biology . On the other side of
the ledger, the compensatory explanatory payoff from the hypothesis
seems meager. If chimpanzees or monkeys do not have minds , then
321
322 ThePhilosophy
of Science
neurobiology and neuropsychology probe the mechanisms and functions of the brain , a reconfiguring of categories can be predicted .
The second question to be asked of the substance dualist concerns
how his hypothesis explains the phenomenon , whatever that phenomenon is. How is it that the nonphysical mind yields unity of
consciousness? How does it unify experiences occurring at different
times? If the answer is that the nonphysical mind unifies because the
experiences are experiences of one substance , then that answer is also
available to the materialist , who can say that the experiences are
The two primary foci for the dualist 's conviction are the logicalmeaningful dimension of cognition and the qualities of consciousness. The importance of these matters has struck dualist philosophers
in different ways , with the consequence that some have gravitated to
one
focus
and
some
to the
other
. One
group
has
taken
the
nature
of
they expect that eventually the logical-meaningful dimension will ultimately have a causal neurobiological explanation . For these philoso phers reasoning is not the stumbling block, partly because the idea
that the logical-meaningful dimension of cognition is fundamentally
noncausal is found objectionable . The second group has just the con verse set of intuitions . Like reductionists , they think that ultimately
consciousness and the qualities of felt experience will be explained in
neurobiological terms . But for them , the difficulty of paramount im -
Mental
portance
lies
they
argue
The
in
,
the
are
logical
has
arguments
These
the
mind
strategy
in
tross
Thus
gued
of
in
that
of
argued
for
,
experience
not
irreducibility
,
arguments
to
be
The
of
will
that
hence
The
of
has
substance
dualists
ghostly
substance
though
there
vanced
by
conviction
experience
and
its
general
notion
this
,
relation
conceptual
its
embedding
an
ostensibly
of
a
holds
property
or
frameworks
theory
is
,
the
More
is
theory
namely
those
noncausal
' s
and
T 2
,
with
just
the
case
subjec
the
brain
,
has
makes
an
"
is
of
property
P
the
in
the
two
to
Al
ad
shared
of
to
between
respect
in
the
argument
function
their
.
hypotheses
mental
hold
the
of
qualities
" emergence
of
dualists
of
emergent
specifically
emergent
goes
is
to
subjec
hands
respect
irreducibly
emergent
, fails
the
the
so
as
from
crux
with
of
people
repeatedly
the
brain
being
many
property
the
and
is
the
Section
complications
the
among
explication
rather
in
with
of
experience
an
to
not
emergent
property
-
reducing
mind
argument
contend
dualists
as
The
,
hands
uniquely
argument
whether
that
to
Subjective
quality
are
from
like
,
life
invoked
powerful
the
the
arguments
been
.
nevertheless
.
if
these
antire
the
seemed
has
differences
even
are
the
in
have
in
and
mental
has
it
property
that
and
Since
but
among
Experience
most
nontrivial
properties
character
who
assorted
is
tive
are
of
reductionism
been
have
they
explanation
that
of
is
philos
deriving
,
of
experience
refutation
and
ar
group
Subjective
extraordinary
experience
ance
and
characteristically
consciousness
neurobiological
have
,
It
arguments
aspect
alba
irreducibility
sophisticated
rejoinders
second
subjective
so
standing
tive
to
most
for
general
dimension
and
those
,
the
Dualism
and
casual
- meaningful
amenable
nature
striking
few
The
property
theory
found
experience
logical
Property
be
the
the
of
- meaningful
the
fit
abandon
idea
irreducible
that
examine
examine
the
not
8 .3
will
subjective
then
see
by
section
to
despite
coherent
experience
cognitive
groups
are
removed
next
nature
the
of
inharmonious
ductionist
an
logical
,
.2
reduc
to
the
subjective
is
the
providing
been
Here
biology
retain
with
with
the
albeit
to
sustained
and
323
strategy
finding
has
but
concerned
subjective
respectably
intuitions
cognition
other
in
reductionist
by
the
physics
these
camps
dualism
substance
concerned
two
be
modern
distinct
both
against
substance
philosophers
ophers
8 .4
can
of
support
idea
use
of
for
to
to
States
dimension
useful
each
- substance
Neurobiological
problems
intuitions
problems
to
- meaningful
been
for
dualist
hopeless
Irreducible
insurmountable
reductionist
tionist
States
appear
order
In
reductive
theories
specified
properties
or
by
of
325
327
328 ThePhilosophy
of Science
(A )
(1)
The
qualia
introspection
( 2 ) The
of
my
sensations
are
brain
are
knowable
to
me
by
.
properties
introspection
of
my
states
not
knowable
to
me
by
Therefore
( 3 ) The
states
.
qualia
of
my
sensations
the
properties
of
my
brain
senses
senses
Therefore
of my brain
seems to be this :
(1) a is F
(2) b is not F
Therefore
(3) a # b
Leibniz 's law says that a = b if and only if a and b have every
property in common . So if a = b, then if a is red, b is red, if a weighs
ten pounds , then b weighs ten pounds , and so forth . If a is red and b
is not , then a ~ b. Assuming
and (B) appear to establish the nonidentity of brain states and mental
states . But are their premises true ?
ise (the properties of my- brain states are not known -to-me-byintrospection ) looks decidedly troublesome . Its first problem is that it
begs the very question at issue - that is , the question of whether or
not
mental
states
are identical
to brain
states
. This
is easy
to see when
329
331
brain
and
how
it
works
. That
is , she
comes
to
understand
completed neuroscience that , among other things , explains the nature of thinking , feeling , and perception , including the perception of
colors. (This is all wildly unlikely , of course, but just suppose.)
Now for the argument : despite her knowing everything there is to
know
about
the brain
and about
the visual
system , there
would
still
be something Mary would not know that her cohorts with a more
regular childhood would , namely , the nature of the experience of
seeing a red tomato . Granted , she knows all about the neural states at
work when someone sees a red tomato - after all , she has the utopian
neuroscience
at hand
. What
she would
not
know
is what
it is like to see
.5
there is to know
(3) Sensations and their properties # brain states and their prop erties
332
sense may involve innate dispositions to make certain discrimina tions , for example. If the first premise uses " knows about" in the first
sense and the second uses it in the second sense, then the argument
founders on the fallacy of equivocation .
The important point is this : if there are two (at least) modes of
knowing about the world , then it is entirely possible that what one
knows about via one method is identical to what one knows about via
a different method . Pregnancy is something one can know about by
acquiring the relevant theory from a medical text or by being pregnant . What a childless obstetrician knows about is the very- same
process as the process known by a pregnant but untutored woman .
They both know about pregnancy . By parity of reasoning, the object
of Mary ' s knowledge when she knows the neurophysiology of seeing
red might well be the very same state as the state known by her
tomato-picking cohort . Just as the obstetrician does not become pregnant by knowing all about pregnancy, so Mary does not have the
sensation of redness by knowing all about the neurophysiology of
perceiving and experiencing red . Clearly it is no argument in support
of nonidentity to say that Mary 's knowledge fails to cause the sensation of redness. Whyever suppose that it should ?
There is a further reservation about this argument . With the first
premise I take no issue, since we are asked to adopt it simply for the
sake of argument . The second premise, in contrast, is supposed to be
accepted because it is highly credible or perhaps dead obvious . Now
although it does have a first blush plausibility , it is the premise on
which the argument stands or falls, and closer scrutiny is required .
On a second look , its obviousness dissolves into contentiousness,
because the premise asks me to be confident about something that is
too far beyond the limits of what I know and understand . How can I
assesswhat Mary will know and understand if she knows everything
there is to know about the brain ? Everything is a lot , and it means, in
all likelihood , that Mary has a radically different and deeper under standing of the brain than anything barely conceivable in our wildest
fligh ts of fancy .
One might say well , if Mary knew everything about existing neuroscience, she would not know what it was like to experience red, and
knowing absolutelyeverything will just be more of the same. That is
an assumption to which the property dualist is not entitled to help
himself . For to know everything about the brain might well be qualitatively different , and it might be to possess a theory that would
permit exactly what the premise says it will not . First, utopian
neuroscience will probably look as much like existing neuroscience as
modern physics looks like Aristotelian physics . So it will not be just
333
more of the same. Second, all one need imagine is that Mary in ternalizes the theory in the way an engineer has internalized Newtonian
physics, and she routinely makes introspective judgments about her
own states using its concepts and principles . Like the engineer who
does not have to make an effort but " sees" the world in a Newtonian
manner, we may consider that Mary " sees" her internal world via the
utopian neuroscience. Such a neuroscience might even tell her how to
be very efficient at internalizing theories. It is, after all, the premise
tells us, a completeneuroscience.
Intuitions and imaginability are, notoriously , a function of what we
believe, and when we are very ignorant , our intuitions will be correspondingly naive. Gedanken-experiments are the stuff of theoretical
science, but when their venue is so surpassing distant from established science that the pivotal intuition is not uncontroversially better
than its opposite , then their utility in deciding issues is questionable.
Moreover , intuitions opposite to those funding premise (2) are not
only readily available, they can even be fleshed out a bit . How can I
be reasonably sure that Mary would not know what a red tomato
looks like ? Here is a test. Present her with her first red object, and see
whether she can recognize it as a red object. Given that she is supposed to know absolutely everything there is to know about the nervous system, perhaps she could , by introspective use of her utopian
neuroscience, tell that she has, say, a gamma state in her 0 patterns,
which she knows from her utopian neuroscience is identical to having
a red sensation. Thus, she might recognize redness on that basis.
The telling point is this : \\rhether or not she can recognize redness is
clearly an empirical question , and I do not see how in our ignorance
we can confidently insist that she must fail . Short of begging the
question , there is no a priori reason why this is impossible . For all I
know , she might even be able to produce red in her imagination if she
knows what brain states are relevant . One cannot be confident that
such an exercise of the imagination must be empirically impossible .
To insist that our make-believe Mary could not make introspective
judgments using her neuroscience because
mental qualia are not iden tical to brain states would , obviously , route the argument round in a
circle.
How could an alchemist assesswhat he could and could not know
if he knew everything about substances? How could a monk living in
the Middle Ages assesswhat he could and could not know if he kne\\r
everything there was to know about biology ? He might insist , for
example, that even if you knew everything there was to know about
biology , you still would not know the nature of the vital spirit . Well ,
we still do not have a complete biology , but even so we know more
, how
it can
be defended
Concluding Remarks
There is a tendency to suppose that a dualist thesis , of either the
substance or the property stripe , can be defended simply by noting
and acknowledging that humans have an introspective capacity , that
we are conscious , and that we are aware of our experiences . But it is a
reduction
would
fall
on
the
retentive
end
of
the
inter
as now conceived. Nobody said light did not exist after the reduction
of optics to electromagnetic
make
perfectly good sense to talk about mental states causing brain states,
since mental states turn out to be states of the brain (cf. Sperry 1980,
Eccles 1977) . Nor is a special notion of causation needed , as Sperry
(1980) suggests .
and
their
nature
are
revised
to some
considerable
of mental
extent
. On
. Rather , he expects
misconceive
and mis -
Mental
understand
those
theory
knowledges
the
terms
of
thusiastically
propriety
geneticist
will
netics
as
It
is
ence
. In
that
will
of
simply
in
,
by
the
and
mental
are
in
remains
the
not
have
degree
8 .4
Intentionality
to
nature
of
physical
defy
now
Eleusis
the
far
too
tell
trans
be
now
early
to
that
predict
aspirations
certainly
of
cannot
intertheoretic
tell
reduction
set
of
the
not
re -
framework
of
character
view
and
These
powerful
will
within
of
dimension
.
a
psychology
phenomenological
point
, but
that
arguments
Intertheoretic
has
on
the
mental
are
of
fact
that
processes
the
focus
of
the
system
or
find
also
of
and
may
can
in
create
beyond
know
gone
degree
,
as
. . and
when
hypotheses
the
for
and
as
.
our
intercon
continuing
that
evidence
relevant
that
mathematical
theory
the
circle
the
such
the
rational
of
things
things
, logically
rule
has
about
, about
construct
the
that
capacity
just
squaring
a coherent
such
.
the
not
for
some
life
representational
impossible
inferences
13 ,
mental
. The
and
) form
data
. We
our
thoughts
methods
draw
in
planets
we
about
them
1, 2 , 3 , 5 , 8 ,
the
dead
even
and
of
have
things
and
of
entirely
can
are
patterns
aspect
neuroscience
appeared
about
subset
We
only
human
humans
1 ,
the
machines
some
Reduction
within
existed
( or
series
not
also
, but
yet
movements
hypothesis
, still
logical
is
Furthermore
nected
neurosci
\ Ted , will
we
considered
arguments
on
explanation
motion
thoughts
concei
Can
one
that
semantic
,
thought
never
perpetual
proofs
be
subjective
and
machine
and
to
not
the
awareness
seemed
ge
and
newly
no
an
Conscious
have
focus
on
states
section
that
of
transmission
- evolved
is
part
reductive
hypothesis
. These
some
next
here
there
dualism
experience
it
earlier
phenomena
defend
neuroscience
property
think
the
remarked
the
that
to
? I
beyond
en
Nevertheless
arguments
duce
as
observing
cards
wheel
will
molecular
- evolved
ac in
he
much
phenomena
forever
, even
neural
the
more
he
and
way
in
much
to
phenomena
to
turn
reduce
and
a necessary
same
reduction
even
to
335
phenomena
,
research
eventual
that
mental
is
neuroscience
an
fail
respect
the
the
meantime
available
as
In
richer
the
mental
is
support
of
In
about
9 .)
States
better
.
research
chapter
will
psychology
theory
part
case
be
talking
possible
with
that
of
enthusiastically
psychology
emergent
is
( See
necessary
empirically
formed
forthcoming
psychological
1!' eduction
Neurobiological
that
be
psychological
support
to
and
will
whatever
eventual
Irreducible
phenomena
comprehensive
this
States
explains
the
the
contradicts
to
explaining
a
an
, such
as x believes
that
P , x wants
that
P , x thinks
that
337
P, x
approach to intentional -
ity has been to see questions concerning IIaboutness" and the objects
of thought as absorbed into the more general question concerning the
contentfulness or meaningfulness of representational states. In those
debates between physicalists and dualists where intentionality has
been the issue , the division concerns whether or not it is possible for
physical states to have content and whether or not physical states can
concerning how thoughts have meaning and how elements of language have meaning attracted an immense research investment
from
philosophers . Matters are far from settled, but it appears that the
network theory of meaning has emerged from the smoke as the most
powerful and comprehensive theory (see below I this section). An
important feature of this theory for our purposes is that , though
compelling on independent grounds , it turns out to be compatible
with physicalist theories of mental states and processes . Actually , it is
compatible with both dualism and physicalism , but becauseit is compatible with physicalism it has been important in showing that the
meaningfulness
results
has
taken
a number
of very
different
forms
. One
line ,
whose failures I have already spelled out in section 7.3, denies that
rationally
meaning and logic in virtue of which it is thought that the capacity for
thought is fundamentally nonphysical and emergent. Popper' s argument closes in on the semanticdimension of propositional attitudes
and on the abstractnature of their propositional objects.
Popperand the World of Intelligibilia
The crux of Popper's argument against reductionism depends on his
idea that there exists a world of abstract, nonphysical objects with
which we interact when we reason, discover a proof for a theorem,
find consequences for a physical theory , use language, think about
arithmetic or quantum mechanics or Godel's incompleteness results.
He calls this realm of abstract objects " World 3," and its denizens
include arithmetic objects such as the integers, the irrational num bers, and the relations between them , mathematical objects, logical
objects , and relations between them , scientific theories , the as-yet -
undiscovered proof for Goldbach's conjecture, and the as-yetundeduced consequencesof theories in physics, neuroscience, and so
forth . It also contains some " embodied" objects such as books and
musical scores. Popper calls the physical world that conforms to
physical laws " World I ," and he claims that mental events and processesbelong to a distinct " World 2."
Popper argues that World 3 exists as an autonomous reality independent of ideas in minds , dispositions to verbal behavior , wiring in
the brain , or any such . That is , World 3 is part of neither World 2 nor
World 1. His reasoning here depends on the view that the laws of
logic cannot simply amount to empirical generalizations about how
persons reason, or even about how professional logicians reason.
These laws of logic are norms or standards concerning what ought to
be, and, by Popper's lights , the validity of these laws does not derive
from the fact that logicians accept them . Rather, logicians accept these
laws because they are valid , and their validity is an independent
matter of World 3 relations among World 3 objects . Moreover , certain
inferences
are
valid
whether
or
not
anyone
has
ever
drawn
those
does
this
bear
upon
the
issue
of intertheoretic
reduction
? In
the following way . Popper aims to show that World 2, the realm of
mental states and processes, cannot be a subpart of World I , and the
argument boils down to this :
Mental
( 1 ) World
when
3 can
States Irreducible
bring
about
uses
a theory
someone
converts
chemical
(2 ) However
through
cannot
are
bring
about
processes
"
( 3 ) World
states
Popper
tence
realm
its
my
objections
his World
( 2 ) that
issue
. For
thing
not
yield
claim
well
, which
technology
ical
and
logical
readjust
wired
their
up
physical
rules
and
not
but
, and
claim
accordingly
, and
the
that
theories
we
show
for
tech -
way : theories
.
exam -
understand
, follow
learn
at
can -
not
are positive
very
mathemat
from
, computers
a sequence
question
of physical
does
this
equations
by going
intelligibilia
sort
of brains
instances
against
is necessary
there
I
3,
in premise
the
. That
workings
namely
such
right
mediation
, solve
the
begs
' s claim
that
find
though
arguing
with
thought
because
whose
through
on
the
exis -
of a World
of thought
that
the
I do
is the claim
simply
Popper
powerful
in some
mental
is to deny
puzzling
can so interact
only
state
programmed
that
. Though
depend
, perhaps
theorems
states , it goes
strategy
can interact
the
systems
prove
laws
do
without
more
physical
can
my
the mediation
would
is even
of purely
only
intel
with
defense
to me in error
, maybe
without
implies
argued
is nonphysical
The
ples
not
is that
assemblies
. A physicalist
cannot
because
is incompatible
be
seems
has
technology
thought
nology
as they
. These
nonphysical
1 , which
' s only
nonphysical
problem
Popper
neural
yield
that
all
not
' s claims
, what
something
3 . One
the
to Popper
only
of World
except
processes
processes
only
objects
a reductionist
assuaged
3 . Rather
world
thought
with
of World
states
will
( rules
perplexity
interact
that
, but
is , abstract
physical
physicalism
that
. That
notions
my
that
3 and
reality
normative
find
be part
be physical
is confident
of World
1 directly
2 . That
by
example
engine
be physical
can
1 , as for
3.
2 cannot
cannot
in the
339
States ?
a steam
World
World
2 cannot
ligibilia
of World
Therefore
:
energy
manipulated
nonphysical
in World
produce
affect
of
changes
and
in World
something
to
3 cannot
intervention
" grasped
changes
to mechanical
, World
the
to Neurobiological
the
past
and
. A computer
is
a sequence
of
through
of functional
states , such
as drawing
a deductive
inference
from
two premises
. By " functional
state " I mean
a state that is characterized
in terms
of its causal
rela tions
with
states
. ( See
scribed
and
functional
in performing
input
also
states , output
chapter
explained
level
at the
in terms
logical
states , and
9 . ) The
level
of what
inferences
machine
of the
other
circuitry
it is computing
, computers
internal
' s operations
interact
functional
can
be
or alternatively
( Dennett
with
de at a
1978a ) . If ,
abstract
ob -
340
341
343
that
he has , and
in turn
the
content
of his beliefs
is a function
of
is a function of
the role that expression plays in his internal representational economy - that is, of how it is related to sensory input and behavioral
and
such
that
the
network
of formal
and
material
economies
may
be
so different
that
translations
are
not
possible.
Meaning is therefore relational in the sense that what an expression
means is a function of its inferential /computational role in the person ' s internal system of representations , his cognitive economy . This
is not to say that an expression has meaning only if someone inter prets or translates it as having a particular meaning . However , it does
imply that isolated expressions do not somehow sheerly have meaning and that mentality cannot somehow magically endow an utter ance with intrinsic meaning . What it does deny is that meaning is an
intrinsic
is
to
be
identified
with
the
state
of
affairs
that
makes
it
. l0
ior is very like my own , be said to have thoughts with meaning and
more generally to have states that represent that p? Now in order to
simulate my outer behavior as closely as the premise asserts , the
robot will have to have an internal system of representations and
computations of a richness roughly comparable to my own . Consequently , it will have elements whose roles have a pattern comparable
to the roles played by elements in my internal economy. But if the
elements in its internal economy are close analogues of my own , if
345
their roles mirror those in my economy, then what else do they need
to have meaning ?
To refuse to assign meaning - meaning as genuine as it gets - to
the robot 's internal states would therefore be to apply a double standard , arbitrarily and to no useful purpose . To bridle here looks like
dogmatism . What the robot means by an expression will , as with me,
be a function of the role that expression plays in the internal representational economy of the robot . If I can find a mapping
between its
elements of its representational economy objectively bear the inferen tial /computational
whether
I encounter
robot
or not .
If the robot turns out to be a fake, inasmuch as its effects are really
produced by a small boy hiding inside , then the intentionality is
derived , for the robot has no system of representations
virtue
of which
inferences
are drawn
and
so forth
. On
of its own in
the other
hand
the
robot
looks
and
smells
different
from
a human
, that
" brain " is a structure of silicon pico -chips , is in the end irrelevant
its
to
intelli -
346
347
Chapter
Functionalist
Further information
and mental activity
of brain processes .
Psychology
J. Z . Young , 1978
9 .1
Introduction
arguments
against
the reduc -
theme
of
folk psychology
that are fundamentally
correct for characterizing
mental states ; (2 ) these categories delimit intentional states and logical
processes , and they will figure essentially
in both the research and
the evolving
theories of a science of the mind ; (3) these categories ,
essential to the psychological
level of description , will not reduce to
categories at the neurobiological
level of description
The implication
for a scientific
psychology , or at
trivial
domain
of a scientific
psychology , is that
neurobiology
can be expected . The implication
for
relevant
domain
of scientific
largely irrelevant
to discovering
processing
at the psychological
In most
variations
on this
.
least some non no reduction
to
research in the
psychology
is that neurobiology
is
an adequate theory of information
level .
antireductionist
theme
the
computer
metaphor
is prominent
, for a number of reasons . One is that a fairly
clear sense can be given to the notion of levels of description
in the
case of computers . The machine can be considered
to have three basic
levels of description
: the semantic level , the syntactic level , and the
level of the mechanism . At the semantic level the machine can be
described
certain
goals , computing
a square
Functionalist Psychology
351
core
defined
idea
in terms
of
functionalism
of their
abstract
is the
causal
thesis
roles
that
""rithin
mental
the
states
wider
are
infor
pain and about what will bring relief , and so forth . The characterization of having the goal of, say, finding a mate will follow a similar
pattern : the goal state will be connected to a complex range of beliefs
and desires , will prompt
namely the lifting of the valves, and it might be instantiated in vari ous physical devices, such as a rotating camshaft or a hydraulic
device. More humbly , " mousetrap" is a functional kind , being
implementable in all manner of physically different devices: spring
traps, assorted cage traps, a sack of grain falling when a trip line is
352
to functionalism
tronic IIgubbins ." The functionalist theory is thus as roundly physi calist as it can be, yet despite their adherence to physicalist principles ,
functionalists have typically rejected reductionism and ignored
neuroscience. Why ?
Plainly , it is not because functionalists suppose that mental states
have no material realization . Rather , it is because they envision that
types of mental states could have too many distinct material realiza tions
for
a reductive
mold
to fit . As
functionalists
see it , for
a reduc
type of physical state, but , they argue, the identities are not forthcom ing . The reason is that one and the same cognitive organization might
be realized
entails
that
or embodied
there
cannot
in various
ways
in various
stuffs , which
be one - to - one relations
between
functional
putational organization of a computer executing a program : computational processesare logical, or at least semantically coherent, and they
operate on symbols as a function of the symbol's meaning, not as a
function of its physical etiology in the machine, and the same program can be run on different machines (Putnam 1967, Pylyshyn
1984). There is nothing in the specification of a cognitive organization , the functionalist
will remind
353
Functionalist Psychology
tion -processing innards are not neurons but microchips , and its cognitive organization
cannot therefore be identical to a particular
neuronal organization , since neural stuff it has not got . Instead , its
cognitive economy will be instantiated in electronic stuff . As we shall
see, the plausibility of these thought -experiments depends on a crucial and highly suspect assumption - namely , that we know at what
level the biology does not matter .
Fictional examples are not really needed to make the point anyhow ,
since there are certain to be neural (structural ) differences between
, it is continued
, there
may
be nontrivial
differences
in
or a street
vendor
. Indeed
, on different
occasions
different
structural
.
'
(Fodor 1975) .
kind . Apart from its implications for the theory that mental states are
identical with brain states, the argument has been deployed to
methodological
If mental
states
and
processes
way .
are functional
kinds
, then
to under
stand how cognitively adept organisms solve problems, think , reason, and comport themselves intelligently , what we need to
understand
is their functional
not going to reveal the nature of the functional organization , but only
something about the embodiment of the functional organization and just one sort of instantiation
scheme
, and
to
this
extent
it
is
removed
from
the
level
of
and
to
this
extent
neuroscience
has
obvious
medical
significance, but structural theory will not enlighten functional hypotheses and functional models . To put it crudely , it will not tell us
how the mind works . Cognitive psychology , in contrast, is focused at
the appropriate level of description , and in cooperation with research
in artificial intelligence it constitutes the best strategy for devising a
theory of our functional
argument .
cognitive
scientists will
figure
business need be known by the cognitive scientist- or the philoso pher , either - since the way the functional organization is instantiated in the brain is a quite separate and independentmatter from the
Functionalist Psychology
355
Functionalist Psychology
357
structures . Now , however , this looks like a merely verbal recommendation about what to call reductions in cases where the predi cates in the reducing theory are relativized to certain domains (cf.
Cummins 1983). As such, it implies nothing about the derivation of
one theory from another or about the autonomy of the sciences. No
grand methodological strictures about what is and is not relevant to
the " functional " theory will be in order . As a merely verbal recommendation it is not especially objectionable, but it has no obvious
utility either .
Dialectically , it does the functionalist no good to deny reduction in
thermodynamics , for then he loses the basis for saying that psychology is on an entirely different footing from the rest of science (En~
1983). After all, if psychology is no worse off than thermodynamics ,
then reductionists can be cheerful indeed . At any rate, the requirements for the reduction of psychology should not be made stiffer than
those for intertheoretic reduction elsewhere in science. (See also
Richardson 1979, Paul M . Churchland 1984.)
The main point of the example drawn from thermody ...~mics is that
reductions may be reductions relativeto a domainof phenomena
. Though
this is called " multiple instantiability " and is draped in black by the
functionalist , it is seen as part of normal business in the rest of science. By analogy with the thermodynamics example, if human brains
and electronic brains both enjoy a certain type of cognitive organization , we may get two distinct , domain -relative reductions . Or we
may, in the fullness of time and after much co-evolution in theories,
have one reductive account of, say, goals or pain in vertebrates, a
different account for invertebrates , and so forth . In and of itself , the
mere fact that there are differences in hardware has no implications
whatever for whether the psychology of humans will eventually be
explained in neuroscientific terms, whether the construction of PS)'chological theories can benefit from neuroscientific information , and
whether psychology is an autonomous and independent science.
That reductions are domain -relative does not mean they are phony
reductions or reductions manque, and it certainly does not mean that
psychology can justify methodological isolation from neuroscience.
En~ (1983) draws a further point out of the thermodynamics case.
Two volumes of a gas might have the same temperature , but the
distributions of velocities of their constituent molecules will be quite
different even while their mean value is the same. To be consistent,
functionalists should again deny reductive success to statistical mechanics since, as they would put it , temperatureof a gas is differently
realized in the two cases. If , on the other hand , they want to concede
reduction here but withhold its possibility from psychology , they
358
need
to
do
species
more
or
If
it
turns
relative
of
on
its
feet
in
on
the
and
have
terization
of
of
ences
As
ately
associated
The
some
the
of
this
are
by
arguing
In
so
neuronal
level
and
they
of
will
chological
Levels
of
There
can
be
example
In
word
of
trilevel
which
be
theory
that
and
more
so
model
suitable
brains
beyond
dispute
to
That
Indeed
is
no
one
no
to
out
con
reduc
between
the
computational
level
of
and
arbitrary
psy
that
statistical
the
mechanics
psychology
is
the
special
antireductionist
levels
of
an
than
learning
will
of
more
is
the
could
will
what
fine
retrieval
is
mecha
also
of
all
could
the
suitable
labor
to
is
knowledge
take
grained
that
division
take
ana
assumption
computers
some
network
theory
be
mechanism
Helmholtz
interactive
which
be
'
organization
during
how
grained
Bacon
since
theory
but
rule
occurring
Neumann
since
one
of
of
coding
should
predicates
instantiations
case
controversial
Von
there
sci
intertheoretic
of
hand
diverse
of
What
at
set
learning
postulates
forth
differences
in
the
reducing
theory
fine
and
inappropri
point
of
changes
subserve
Brain
be
lives
basic
if
general
uncontroversial
than
will
more
the
functional
the
Mind
mapping
case
that
is
blue
it
func
one
one
thermodynamics
synaptic
grained
structures
organic
province
radical
the
must
and
fine
tomical
nisms
cellular
more
the
there
the
profuse
the
that
to
multiplicity
claim
in
that
learns
they
deal
appreciation
than
good
so
to
true
charac
typically
of
to
the
be
likened
Organization
is
theory
that
be
psychology
general
of
and
out
very
the
point
explanation
point
cannot
one
concede
will
core
entirely
functional
the
is
details
they
predicates
case
be
the
is
frequently
the
from
predicates
to
the
it
as
of
been
wish
is
and
and
"
onto
that
arguing
psychology
on
stands
abstract
Functionalism
package
package
that
it
their
can
functionalism
has
may
'
whole
theory
Antireductionists
tinue
of
conditional
level
of
nonreducibility
for
and
system
matter
to
falls
terms
for
do
Functionalists
another
the
section
not
higher
is
term
the
Nothing
processing
argue
with
point
'
degree
result
reductions
the
states
in
domain
this
instantiability
in
reduction
to
multiple
reducibility
however
does
functionalism
naysaying
mental
autonomy
from
identified
information
wished
reduction
what
kinds
functionalism
question
breathes
tionalists
between
argument
are
wider
functional
argument
without
get
antireductionist
states
makes
functionalists
to
neuroscience
are
the
the
the
differences
enough
to
mental
that
neutral
lucky
kinds
after
roles
conception
tion
are
mental
that
causal
of
we
of
own
hardware
independent
thesis
predict
psychology
that
is
The
that
human
thesis
thesis
merely
individuals
out
the
than
between
also
as
even
all
his
of
Functionalist Psychology
neuroscience
359
the level of intracellular dynamics . (See also Lycan 1981a.) And even
if there were just three, neurobiological theory (see chapter 10) challenges that way of specifying their organizational description . How
many levels there are, and how they should be described, is not
something to be decided in advance of empirical theory . Pretheoretically, we have only rough and ready- and eminently revisablehunches
J
about
what
constitutes
a level
of or2"-: anization
circuit
, the
behavior
. And
within
each
level
further
substrata
can
affair
, or
if
some
cell
assemblies
or
modules
have
a transient
more or less fine -grained . The cellular approach taken by Kandel and
his colleagues (Hawkins and Kandel 1984) showing modification in
presynaptic
neurotransmitter
release in habituation
at a lower level than studies by Lynch and his colleagues (Lee et ale
1980) showing modification of synapse numbers and synaptic morphology correlated with plasticity in behavior , which in turn is at a
(slightly ) lower level than the studies by Greenough and his colleagues (Greenough , Juraska, and Volkmar 1979) on the effect of
maze training on dendritic branching . We then ascend to the mul ticellular studies in the hippocampus done by Berger, Latham, and
Thompson (1980), and from there up (a bit ) to the cell assembly stud-
Functionalist Psychology
361
the artificial units would have to have both action potentials and graded
potentials , and a full repertoire of synaptic modifiability , dendritic
growth , and so forth , though unlike neurons they might not need to
have , say , mitochondria
or ribosomes
. But , for
all we
know
now
, to
the
physical device- the brain itself - will not produce theoriesof func tional organization . Now we have already seen that the functional structural distinction will not support the simplistic idea that
psychology does functional analysis and neuroscience does structural
analysis , and that there are bound to be many levels of organization
between the level of the single cell and the level at which most cognitive psychologists work . It is important as well to emphasize that
when neuroscientists do address such questions as how neurons
manage to store information , or how cell assemblies do pattern recognition , or how they manage to effect sensorimotor control , they are
addressing questions concerning neurodynamics - concerning infor mation and how the brain processes it . In doing so, they are up to
their ears in theorizing , and even more shocking, in theorizing about
representations and computations . If the representations postulated
are not sentence -like , and if the transformations postulated do not
resemble reasoning , this does not mean the theory is not functional
" func -
that
we
know
what
level
is functional
and
what
naive to
is structural
362
Functionalist
Psychology 363
My
guess
is
thought
that
it
What
at
clear
,
an
search
the
case
of
The
of
cially
of
Co
macro
the
the
First
the
the
basis
laws
For
statistical
which
versely
work
now
statistical
theory
,
The
genetics
instances
biology
re
using
indicate
why
synopsis
e .g
to
to
and
the
the
the
the
match
equilibrium
"
it
of
can
thermodynamics
this
More
form
generally
eventually
processes
generally
conceptual
may
injected
them
into
traditional
Con
process
statistical
be
among
paradox
.
a
into
differences
- equilibrium
to
theory
" back
that
the
led
to
ergodic
undergoing
Gibbs
and
attempt
entropies
its
has
that
itself
various
hope
non
co
between
of
retaining
some
the
construct
thermodynamical
transform
is
both
properties
the
of
while
en
at
mechanics
attempt
entropy
injection
.
Our
and
of
discrepancies
is
to
,
,
made
statistical
the
the
to
solution
there
equilibrium
swift
development
the
,
espe
.
ap
allow
as
well
value
infectious
of
Then
were
of
by
entropies
the
extension
: 49
of
and
and
and
' s
thermodynamics
afoot
proper
physics
Gibbs
to
the
explain
was
thermodynamics
for
thermodynamical
through
basis
paratus
( 1981
it
constructs
is
its
the
into
shall
displays
discoveries
Hooker
flavor
quantities
enrichment
ing
however
from
thermodynamical
and
led
mechanical
"
the
example
mean
shall
prac
development
temperature
precisely
entropy
values
In
gain
mechanics
development
of
issue
concordant
corresponding
development
at
co
- evolution
ideology
as
levels
gives
the
Boltzmann
" back
of
influenced
for
fields
statistical
readjusted
mathematical
heavily
co
learning
of
nature
micro
two
.
to
follows
the
- evolutionary
benefits
and
and
of
been
and
co
enriching
what
that
Development
mutual
expanded
evolution
of
the
and
of
was
the
memory
reveals
nothing
namely
the
experiment
science
has
In
philosophical
enough
mutually
- evolutionary
the
understanding
tropy
on
thermodynamics
well
of
outside
prefer
other
not
discipline
examples
should
case
ideology
research
Benefits
history
methodology
on
many
and
been
as
research
drawing
the
typically
isolationist
psychology
The
has
with
thought
that
psychology
competing
flaw
much
is
theories
least
adopting
the
is
of
tice
: too
think
evolution
lar
shares
- experiments
co
- evolution
chemistry
disease
and
more
and
,
and
discoveries
can
,
of
of
also
recently
of
immunology
one
of
in
the
dynamics
classical
often
genetic
provoke
development
,
genetics
nonequilibrium
and
level
seen
and
microbiology
at
be
astronomy
of
of
the
theory
and
of
molecu
thermodynamics
engineering
further
In
experiments
these
further
corrections
questions
It
may
it
tion
The
conflicts
with
trouble
reasoned
1974
This
and
In
worth
evol
genetics
mutation
why
there
causal
structural
The
Looked
at
some
some
by
seem
of
the
that
same
instead
of
tionally
the
tion
As
gene
1982
one
There
is
not
cause
explain
genetics
as
it
"
There
molecular
phenotype
and
is
traits
In
no
there
is
it
has
many
the
been
indefinitely
sug
many
rela
biology
the
func
and
molecular
of
again
short
as
genetics
to
biology
conversely
genes
perhaps
did
structure
surround
And
the
what
the
phenotypic
many
that
but
the
were
Not
classical
it
later
biochem
between
and
many
trait
in
specified
be
gene
that
matters
in
of
what
the
differences
genes
find
as
reduction
structurally
to
to
differences
relation
their
functional
were
biochemical
one
as
puts
or
those
who
have
the
molecular
to
of
that
distinct
the
set
and
that
genetics
only
chromosome
frustrate
input
of
identified
of
in
certain
expectations
explain
of
genetic
the
material
"
Importantly
hopeless
to
terms
was
be
mechanisms
the
Kitcher
is
evident
yield
bad
circumstances
genes
so
be
is
given
gene
became
to
molecular
be
analogies
in
expression
in
between
Philip
: 357
It
the
could
and
may
well
Hull
theory
early
result
coveted
of
it
not
is
transmission
the
of
matters
material
gested
The
discovery
base
relation
genetic
biochemical
specified
many
always
divergent
especially
the
genetics
thought
not
DNA
to
See
reduc
there
hereditary
in
way
many
the
compensating
the
owing
to
could
DNA
was
did
of
molecular
that
was
to
of
itself
DNA
owing
this
to
discouraging
of
role
organism
contributed
complexity
units
segment
and
conditions
traits
trait
complex
ical
so
essentially
in
the
strip
exceedingly
show
has
and
neurobiology
phenotypic
identifiable
identifiable
lead
because
Their
for
in
phenotypic
be
to
genetics
do
specified
problem
particular
would
are
mechanism
for
the
output
they
properties
properties
are
and
will
basis
and
and
wish
tion
given
molecular
look
provoke
genetics
what
evolution
will
closer
recombination
is
that
psychology
genes
and
fertilization
that
of
classical
structural
never
case
Co
classical
is
the
it
is
turn
of
co
to
in
of
that
"
which
explorations
example
said
gene
that
and
Reduction
be
II
bad
idea
will
new
example
the
it
level
for
the
fact
reducing
case
it
to
in
skepticism
tween
other
ideas
that
is
in
the
and
objected
genetics
succeeded
be
molecular
because
at
corrections
input
that
have
not
to
output
reduction
claim
that
phenomena
transmission
genetics
rejected
wished
described
genetics
is
irrelevant
in
molecular
is
to
transmission
an
genetics
as
genetics
by
autonomous
transmission
science
genetics
can
On
or
the
365
Functionalist Psychology
contrary
the
they
reverse
significance
the
of
molar
This
is
program
in
one
- to - one
and
of
from
not
the
like
an
and
reconfiguration
of
. ( See
reasonable
( unitary
that
functions
onto
unitary
system
complex
between
Consider
a set
of
frequency
together
ity
is
in
a suitable
tors
momentarily
compensated
absorbed
from
an
it , with
property
" has
system
it
the
system
erty
' s being
the
/ output
increased
as
by
a virtual
a whole
irreducible
governor
, but
of
? Yes , and
' s operation
is
a property
of
of
none
entire
produces
system
, once
, it
is entailed
in
phase
and
con
seen
,
a
engineers
" . The
a property
of
. Does
this
mechanism
that
by
en -
with
governor
the
understood
variabil
, genera
functions
control
is
joined
generator
components
is
singly
ha \ ' e their
f"
sort
fluctuations
output
a single
its
compel
this
. Taken
phase
this
system
but
10 % . Taken
average
the
complicated
frequency
in
map
when
, statistically
reliability
no . For
always
which
, " virtual
of
truth
so
with
value
, or , as
powerful
not
of
is a
. However
generators
, as
reliability
, more
reasonable
, makes
but
collective
the
input
the
simple
because
view
is
there
and
is
. The
of
of
of the
be
Hooker
remaining
remainder
frequency
a single
a long
ahead
point
category
' s course
is that
do
to
average
behind
the
momentarily
by
input
greatly
render
for
level
generators
figure
are
a frustrating
G , each
, their
that
fluctuating
express
the
network
of
, generators
ergy
the
is
at 60Hz
some
of
basis
, and
following
power
around
the
, it
by
.
look
, there
mechanism
is
generators
electrical
10 %
there
not
a fragmentation
likely
, described
electrical
a fraction
are
- output
. The
1976
categories
does
. Indeed
effect
especially
output
variability
only
versely
of
input
and
Dewan
psychological
unitary
research
reductionist
reductionist
on
the
of
theory
psychological
fully
the
is , when
that
expectation
that
is
the
predicts
the
natural
that
current
frequency
the
at the
input
alternating
more
find
. This
from
clear .
issue
and
postulated
it
is postulated
to
see
the
units
molecular
possibility
folk
accomplishing
try
the
categories
why
, the
expectation
ling
example
of situation
in
to
as
the
exactly
concerns
neurobiological
show
a system
structures
is
route
is
the
the
for
strattgy
reasonable
much
functional
possibility
onto
, this
of
system
) mechanism
research
so
explore
the
who
psychological
4 . 1 . ) To
of
of
those
the
one
a fragmentation
a property
of
arguing
then
the
units
viewpoint
the
section
profile
by
to
reduction
, I shall
When
to
blocked
of
signs
output
structural
one
in
,
between
importance
map
to
reductionists
disagreement
mappings
reductionist
obstacle
with
the
the
central
will
memory
of
neuroscience
categories
already
agree
heart
theory
issue
Now
typically
The
this
the
of
prop
conjunc
Science
is the
virtual
governor
?" and
so on . The
answers
must
be
Genes, it appears, are more like the virtual governor in the example
than like a distinct component
scribe the situatio]l in genetics, therefore, is to say that there really are
no genes, in the sensespecifiedby such questions as " What thing in the
DNA is a gene ?" or " Where are the genes ?" The assumption
that
in that set .
characterizations
of the sys-
in Philip Kitcher
1982,
Functionalist Psychology
367
368
Functionalist Psychology
369
370
and in Alzheimer 's patients , patients who have undergone electroconvulsive therapy , and patients with tumors or with other lesions
have been intensively studied and compared both at the performance
level and in terms of the underlying pathology . Important differences
have emerged, and the data have suggested to some neuropsycholo gists that there are at least two distinct physiological systems subserving higher long-term learning and memory .
One line of evidence in favor of the hypothesis depends on showing that the two capacities are dissociable, and the most striking example of such dissociation is displayed by Milner 's patient , H .M . At
age 27 H .M . underwent surgery as a means of last resort to control his
intractable epilepsy . The surgery involved bilateral resectioning of the
medial temporal lobe and hippocampal structures (figures 3.3 and
3.5) (Scoville and Milner 1957). In tests following the surgery an unexpected side effect was discovered: H .M . has profound anterograde
amnesia and can recall virtually nothing of any events that have happened since the surgery . As Scoville and Milner have observed, he
appears to forget everything as soon as it happens . He cannot remember what he had for lunch or whether he had lunch , he cannot
recall having met the hospital staff he sees many times a day or how
to get to the bathroom . Nor can he recall being told that his favorite
uncle has died , and each time he inquires about his uncle and is told
of his death, his grief is renewed . His IQ was slightly above average
preoperatively and is still so postoperatively . H .M . has quite good
retrograde memory , especially for events a year or so before surgery,
and he easily recalls events from his early life . Eventually H .M . came
to understand that something was very wrong in his life , and he
commented ,
Right now , I'm wondering . Have I done or said anything amiss?
You see, at this moment everything looks clear to me, but w hat
happened just before? That's what worries me. It 's like waking
from a dream; I just don't remember . (Milner 1966:115)
Despite his severe anterograde memory deficit , it has been found
that H .M . does have a retained capacity for learning certain kinds of
things . For example, Milner (1966) showed that he could learn a mir ror tracing task, and Corkin (1968) showed that he could learn rotary
pursuit and manual tracking . Although these achievements might be
considered merely noncognitive , sensorimotor skills , it is remarkable
that H .M . has also mastered, at a normal rate, the Tower of Hanoi
puzzle (figure 4.1). That is, when presented with the puzzle, H .M .
can solve it in the optimal number of steps, and he can proceed to the
Functionalist Psychology
371
end if started anywhere in the middle (Cohen and Corkin 1981). This
is a nontrivial intellectual feat.
More puzzling still , on each occasion when he is asked to attempt
the puzzle , he does not recall having learned how to do it , and he
does not even recall having encountered it before. That is, he has no
conscious recognition of the puzzle that he can verbally report .
When , having solved the puzzle in the optimal number of steps,
H .M . is asked to explain his expertise, he shrugs it off with a confabulatory explanation , such as that he is good at puzzles, this one is
easy, and so forth . He has. acquired an impressive cognitive -cummotor skill , yet so far as he is able to tell us, the puzzle is always a
complete novelty . Control is therefore also problematic here because
H .M . can initiate and successfully complete an extended, intellectu ally demanding task, even though he has no awareness that he has
the knowledge or that he is executing his knowledge on the task at
hand . If H .M . has no awareness of his skill , how , one wonders , does
the nervous system execute the long sequence of steps necessary for
the puzzle's solution ? The point is, in some sense H .M . is not aware
of what he is doing , yet what he is doing is cognitive , complex, and,
should we say, intentional . To put it crudely , it is as though some
part of H .Mis nervous system knows what it is doing and has the
relevant complex intentions , but H .M . does not . What does H .M .
think he is doing ? Thus the problem of control . (See also Warrington
and Weiskrantz 1982.)
It is this unpredicted dissociation of capacities that has moved some
neuropsychologists to postulate two memory systems, each with its
own physiological basis. In attempting to key onto the criterial features of the lost and the retained capacities of H .M ., Squire and
Cohen (1984) propose that the distinction be made as follows : " descriptive memory " is the capacity to verbally report recollections, .and
" procedural memory " is the capacity to exhibit a learned skill . Al though this way of framing the distinction is useful in some measure,
it is acknowledged by Squire and Cohen to be a preliminary and
imprecise taxonomy , and much remains to be discovered about the
range and nature of each capacity, assuming there are indeed two .
Studies of animal models tend to confirm the fractionation of memory components (Weiskrantz 1978, Mishkin and Petri 1984, .ZolaMorgan 1984), but interpretive questions remain concerning which
anatomical structures explain the observed deficits and the degree to
which animal models illuminate the human cases. Anatomical studies
are certainly important in determining whether there are distinct anatomical structures subserving dissociable memory systems and in
372 ThePhilosophy
of Science
suggesting further ways to test amnesic patients in order to determine just which capacities are retained and which are lost . It could be
that PET and cerebral blood flow scans will help resolve some of the
issues currently in dispute . (For a cautionary discussion concerning
the Squire version of the two -systems hypothesis , see Weiskrantz
(forthcoming ) .)
At the same time , of course, psychologists have collected immense
amounts of data concerning learning and remembering in the intact
human . Learning curves, forgetting curves, distinctions between unaided recall, cued recall, and recognition , and the role of attention , for
example, have been enormously useful and have been employed at
other levels of investigation . Some psychologists have framed a " two systems" hypothesis but have drawn the boundaries quite differently
from the neuropsychologists . Tulving (1972, 1983) distinguishes
episodic from semantic memory (neither of which encompassescognitive skills ), and Kinsbourne and Wood (1975) have used this distinc tion in their research on amnesia. Others (Norman and Rumelhart
1970) have distinguished between memory for item and memory for
context, and this rather different taxonomy has also been used in
interpreting data on amnesics (Huppert and Piercy 1976).
The assorted more-than -one-system hypotheses have provoked
widespread testing of intact humans and have occasioned new studies on
I skills (Kolers 1975, Rumelhart and Norman 1982), the role of
consciousness in the various kinds of memory (Posner 1984, Poulos
and Wilkinson 1984), and so forth , in order to try to discover the
nature of the two capacities, if such there really be, and whether there
is theoretical justification for postulating yet other capacities. What is
striking about the research on memory and learning is its interdisci plinary character. Researchersworking at one level are always on the
lookout for , and are often successful in finding , results elsewhere that
can help them in devising useful experiments or in reconfiguring the
memory -learning taxonomy .
One hope concerning the research in memory and learning is that
the explanation of the biological basis of relatively simple plasticity
such as habituation in Aplysia can be used as a scaffolding for addressing the neurobiology of other kinds of learning and for figuring
out the neurobiology both of habituation and (eventually ) of other
kinds of plasticity in humans . The current classification of kinds of
learning may well be reconfigured even further in unpredictable ways
as the research at all levels continues . There is a sense in which , so far
as higher learning is concerned, we cannot be at all sure yet what the
explananda are. For example, we do not have a principled way to
Functionalist Psychology'
373
374
completely misconceived . Worse again, they may depend on a misconception , deeply entrenched in folk psychology , of the phenomenon to be explained . As neuroscience and psychology co-evolve, both
will need folk psychology less and less. A metaphor that is perhaps
more apt likens the co-evolutionary progress to two rock climbers
making their way up a wide chimney by bracing their feet against the
wall , each braced against the back of the other .
What the mind -brain is doing - even as described at the level of
input -output functions of the system- is not an observational matter,
to be read off simply by looking at the behaving organism . Rather, it
is a deeply theoretical matter . Some initial theory is essential to get
the whole enterprise going , and broadly speaking, folk psychology is
that initial theory . We have already gone beyond folk psycholo~y,
and as neuroscience and psychology co-evolve, the likelihood is that
the initial theory will by inches be revised, lock, stock, and barrel .
(For more specific arguments demonstrating weakness in folk psychology , see section 9.5.)
The co-evolutionary developmentof neuroscience
and psychologymeans
that establishingpoints of reductive contact is more or less inevitable. As
long as psychology is willing to test and revise its theory and hypoth eses when they conflict with confirmed neurofunctional and neurostructural hypotheses, and as long as the revisions are made with a
view to achieving concord with the lower -level theory , then the
capacities and processesdescribed by psychological theory will finally
find explanations in terms of neuroscientific theory . (The same is
true , of course, for the revisions and reconstructions in neuroscience.) If certain capacities specified in folk psychology turn out to be
" virtual governors " so far as neuroscience can determine , then a
coevolutionary strategy will elicit revisionary descriptions of the
capacities, based on what seems pretty certain at the neurofunctional
level . The capacities, newly described, will be explained by the
neurofunctional theory that prompted them . Obviously , this is an
oversimplification of what is a horrendously complex business, but I
am after the general features at this point . The heart of the matter is
that if there is theoretical give and take, then the two sciences will
knit themselves into one another .
How thoroughgoing a reduction will ultimately be achieved remains, of course, to be seen, and a rich reductive integration could be
frustrated by a variety of factors. First are the practical difficulties ,
such as disruption to research- by plagues, wars, and so forth . Second , the problem of understanding how the mind -brain works may
be beyond our finite capacities. The brain, in the event, may be more
complicated than it is smart, in which case reductive consummation
Functionalist Psychology
375
of neuroscience and psychology may forever elude us. One does not
believethis , but it is a possibility to be admitted . Third , and not unre lated, we may simply lack the mathematics adequate to the task. The
sobering example here is the reduction of chemical properties to
quantum mechanical properties , where the general outlines are clear
bility that cannot be ruled out a priori is that mental states have
certain properties not explicable in terms of neurofunctional
tion or neurostructural dynamics .
organiza -
There are, however , disanalogies between the case of light and the
case of mental
states
and
processes
, the
most
salient
of which
is this :
and was dashed by the discovery that electrical properties are also
fundamental . For the case of psychology to be exactly like that , it
would have to turn out that mentality is somehow a fundamentalproperty of matter , and for that there is not the smallest shred of evidence .
argue
the
merits
of co - evolution
of theories
is not
to advocate
is up
to . No
one
could
do
that
even
if he wanted
to . On
strong version of the autonomy thesis, psychology goes its own way ,
and in so doing , it is essentially isolated both in matters of discovery
376
and
in
matters
of
evolutionary
each
be
as
the
for
perception
that
keeps
of
unity
on
theory
as
1974
and
the
sentential
tions
and
lines
its
is
section
shall
argue
seeing
the
tion
cal
rules
tion
be
shall
in
as
logical
followed
between
co
cognitive
the
theory
To
its
excuse
Oppenheim
and
contrary
view
concerns
the
and
it
may
insisted
by
case
be
see
there
are
as
memory
and
computa
neuroscience
and
in
section
compelling
. 6
reasons
problematic
and
evolutionary
fundamental
to
view
that
and
the
reduced
this
integ
computa
conceded
that
nor
for
ideology
for
thus
for
un
claimed
but
causal
of
manipulation
of
on
are
one
on
distinc
to
one
states
in
Psychologi
Psychological
hand
stand
explainable
representations
the
only
on
times
state
not
also
relations
many
relationships
changes
processes
rely
structure
and
it
meaningful
and
functionalists
and
relations
seen
and
for
the
sizes
science
learning
that
meaning
For
the
all
Once
of
favors
Thus
be
program
have
representational
rest
thesis
will
function
we
of
representations
reductive
logical
for
old
Reduction
between
between
is
units
representations
paradigm
and
states
demarcating
showing
the
the
no
of
the
sentential
distinction
run
or
plummets
autonomy
revised
present
it
it
long
coherence
the
it
neither
Representations
another
are
the
dismissing
do
1963b
the
it
preferring
. 5
to
in
against
even
equivocally
In
to
while
the
not
epistemic
credibility
characterizing
science
1980
its
model
nevertheless
sound
hegemony
with
irreducibility
framework
tions
The
neurosci
hypothesis
Churchland
of
logical
redrawn
but
and
be
theoretical
In
absolute
may
capacities
is
assumption
of
1963a
1975
other
in
theoretical
cohere
combat
attention
of
working
for
and
Feyerabend
because
Paul
the
subject
susceptibility
isolation
ideological
worthiness
suspect
de
mainstay
rity
of
but
by
the
such
isolation
as
to
is
hors
1958
Fodor
"
having
ledger
Putnam
is
but
belief
up
from
confirmation
run
or
of
sentences
exempt
advocated
thumping
of
it
co
level
reach
thing
neatness
memory
general
another
is
criterion
from
is
Short
any
on
disconfirmatory
one
quite
puritanical
tub
principal
is
science
of
positivistic
of
is
at
discoveries
in
the
should
converge
to
because
steam
revision
learning
the
principle
sake
on
On
neuroscience
inquiries
open
honest
head
in
The
mistake
science
isolation
the
up
do
be
and
when
from
be
works
And
should
psychology
would
neuroscience
psychology
and
they
each
of
ence
disconfirmation
other
example
isolation
science
to
of
matter
from
of
vulnerable
discoveries
"
disconfirmation
picture
and
This
terms
of
connec
symbol
ma
Functionalist Psychology
377
Functionalist Psychology
379
, cognitive
there
is
terize
the
or
no
relate
sentence
logic
( see
may
be
general
the
hand
shall
take
it
that
This
. This
be
very
to
many
defend
have
their
distress
tive
Autonomy
and
the
to
respectable
-
of
tious
are
go
overextended
beyond
, since
minority
The
between
tween
to
states
neurobiological
be
of
the
specified
defined
by
argument
it
autonomy
of
attitude
psy
unequivocally
on
of
the
sentential
argument
liberal
as
the
in
their
the
work
since
swim
not
wish
Certainly
as
to
it
cognitive
Michael
about
cognitive
, will
therein
territory
example
, assum
position
of
is
possible
psychology
a central
vagueness
will
neuro
Gazzaniga
how
much
of
) .
cogni
show
,
but
or
only
that
the
view
that
claims
have
of
premises
is
have
taking
to
interdependence
heart
the
no
psychology
is
psychology
is
psychol
part
mind
- brain
support
very
,
is
the
,
of
conten
I
-
, and
the
to
say
the
autonomy
arguments
within
the
arguments
obvious
convinced
these
are
autonomy
. That
not
of
works
. What
and
evidently
been
cognitive
cognitive
quarrel
irreducibility
can
reach
that
indispensable
how
I
not
cognitive
necessary
, and
is
cognitive
that
for
is
the
opposite
of
is
the
as
- Rakic
proceed
philosophers
opinion
brunt
case
the
neuroscience
understand
larger
after
be
, for
from
arguments
enough
many
to
describe
their
's
the
argues
themselves
autonomous
what
psychology
there
are
scope
intentionality
and
much
have
autonomy
as
fair
the
these
least
be
coming
are
cases
charac
Assessed
pursuing
they
such
modeled
cognitive
attention
us
neuroscience
worth
enterprise
think
let
from
. Taken
Goldman
Irreducibility
should
irreducible
is
)
be
carving
a certain
stake
arguments
psychology
ogy
at
1984
occupies
Patricia
, granting
is
to
sentential
the
of
who
,
. Thus
off
researchers
science
What
or
hived
example
Nevertheless
in
not
of
that
consider
learning
discipline
( for
defend
unsatisfactory
who
on
those
science
of
should
theory
an
work
a whole
autonomy
attitude
inevitably
who
latter
: content
autonomy
indeed
the
psychologists
but
the
how
sentences
to
suggests
wide
sentential
is
contrast
( 1980
as
psychology
to
of
, if
application
Pylyshyn
psychology
intended
limited
the
in
)
straightforward
meant
but
By
between
is
that
other
cognitive
is
the
it
is
theory
representations
relations
Given
particular
reason
folk
another
theory
that
in
attitude
of
) .
. The
/ logical
the
assumed
in
On
ing
and
below
chology
on
palpable
one
the
psychology
that
to
attitudes
so
even
semantical
they
sentential
, and
( not
relevant
how
by
skills
theory
Indeed
philosophy
.
my
cannot
reply
attacks
be
states
the
explained
. I
central
in
think
this
idea
terms
opinion
that
of
logical
causal
is
relations
relations
fundamentally
be
Functionalist Psychology
381
about
limitations
of neuroscience
. It assumes
either
that
there
is too
hard
to
ever
tell
us
much
. It
therefore
fails
to
recognize that theorizing about information processing in cell assemblies , groups of such assemblies , and so on up , is as much a part
VIew
neuronal arrays. Why should not such a theory explain the logical
and meaningful relations between states at the psychological level?
How , a priori , do philosophers know that it cannot? \'\JThatcan be
382
Functionalist Psychology
383
set of vague notions flying in loose formation . This may be all very
well for a folk theory going about its humdrum business, but if there
is no such thing as the content of someone's propositional attitudes ,
then the project of finding a place for the propositional attitudes in
cognitive psychology , let alone neurobiology , is severely hampered .
The second problem is that some of the criteria routinely invoked
within folk psychology for specifying content rely on features that are
irrelevant to the causalrole of the mental state in interaction with other
mental states, stimuli , and behavior . Noteworthy culprits here are
semantic features- namely , truth , reference, and meaning . Now
such criteria will give completely irrelevant content specifications if
what we seek are psychological generalizations that articulate the
causal role in cognition of beliefs and desires. The brain , as Dennett
(1981) has put it , is a syntactic engine, not a semantic engine. That is,
roughly , the brain goes from state to ,state as a function of the causal
properties of the antecedent states, not as a function of what the
states are " about" or whether they represent something true about
the world . So far as the question of reduction is concerned, this
means that at best, considerable correction and reconfiguration of the
folk psychological concepts of belief and desire are required . It may
be possible, however , that a syntactic analogue
of the sentential attitudes will be able to do duty i-n a scientific psychology and will fit in
comfortably with a cognitive neurobiology in the geometric style (sections 10.3- 10.6). Even so, of course, these newly characterized states
and th ~ relations between them can be expected to have a neurobiological reduction , for th'e reasons just discussed.
Semantical issues are notoriously difficult , and this is not the place
to do anything but take the lid off the jar and let a few odors escape.
Suffice it to say that Stich's breakthrough on the question of semantics bears importantly on the possibility of smoothly reducing folk
psychology to neurobiology
- - , as well as on how representational
...
states are to be understood in cognitive science. If Stich is right , and I
think he is, then it is already clear that the propositional attitudes ,
qua folk psychological categories, are coming apart . Therefore, when
antireductionists parade these categories in all their folk psychologi cal regalia as irreducible , the irony is that it is their lack of empirical
integrity that prevents their reduction .
RevisionaryReductionsand Folk Psychology
If the reduction is smooth , relatively little revision of the top-level
theory may be required ; if it is bumpy I conceptual revision may be
greater. Notwithstanding reduction and revision , belief talk , like
gene talk , could continue to have a useful role- and not just after
Functionalist
Psychology 385
in , and the generalizations that are true of them are so by virtue of
their manufacture . By contrast, human brains were not literally built
to instantiate the generalizations of folk psychology , and what
generalizations do truly describe our inner processes is a matter for
empirical research . In the case of evolved organisms , ourselves and
others , we must find out what generalizations are true and what
in many
dimensions
, and under
pressure
from
discoveries
at
the
value
of co - evolution
of theories
revolu tionary ?
386
Functionalist
Psychology 387
of an isomorphism holding between the states of the person (his
brain ) and the relevant sentences of a set.
2. The theoretically important relations between cognitive states
are characterized by means of the resources of logic . These obtain
in virtue of the aforementioned isomorphism . (Various theories
of logic will be variously favored by cognitive scientists.)
3. The transitions between states are a function of the logical
relations holding bet\\reen the sentences identifying the states,
which in the most straightforward case will consist of inference,
abductive and deductive . Again , by virtue of the isomorphism .
4. The evaluation of the cognitive virtue (rationality ) of a system
is a function of the extent to which it succeedsin doing what the
favored theory of state transition (i .e., theory of logic) says it
ideally should do.
Any sentential theory needs a theory of how it is that internal states
have content, since it is plainly ad hoc to suppose it a lucky accident
that the relevant isomorphisms systematically obtain . Fodor's theory
(1975, 1981) is that sentential attitudes (alias propositional attitudes )
are relational states, where one of the relata is a sentence in the
organism's languageof thought. Thus, if Smith has the thought that
gulls' eggs are delicious , then he stands in a certain relation to a
sentence in his language of thought , namely the sentence whose
English translation is " Gulls ' eggs are delicious ." Broadly speaking,
the information processing is manipulation of representations, and
representations are symbols of the language of thought . Thinking is,
to put it crudely , sentence-crunching , and the machine analogy
underwrites the intended sense of symbol manipulation .
Within cognitive science there is considerable loose talk about cognitive representations , where the question of the nature and status of
representations is left conveniently vague. Fodor's theory makes explicit what is an implicit commitment in sentential/representational
hypotheses generally . There are, it should be emphasized, many cognitive psychologists who do not adhere to the sentential theory and
who investigate cognitive capacities while awaiting the development
of a nonsentential theory of representations .
The foregoing is just a thumbnail sketch of the main features of a
sentential theory of information processing, but the finer points and
the in -house disputes can for now be set aside.
Cognition and Sentence
-Crunching
How plausible are the frame\l\rork assumptions of the sentential
paradigm ? The theme of the ensuing section is that as an approach to
Functionalist Psychology
389
Short of conceding that there is substantial nonsentential representation , two solutions are possible. (1) One can argue that behavior in
the infralinguistic organism is not really a cognitive product . Language learning by the toddler , on this view , is not in fact cognitive in
the intended
in
for supposing
out
concerns
how
the
infant
uses
for
Mentalese
to learn
the
lan -
he can only acquire English concepts for which there are Mentalese
correlates. This means that there is no such thing as real concept
learning , in the sense that wholly new concepts are added to the
conceptual repertoire .
atom, force field , quark , electrical charge, and gene- are lying readymade in the language of thought , even of a prehistoric hunter gatherer . It is difficult
there is something like a language of thought in linguistically accomplished humans . The concepts of modern scienceare defined in terms
of the theories that embed them , not in terms of a set of " primitive
1980b.) If it
Functionalist
Psychology 391
the opposition ' s rally and hears an Ottawa Liberal say, " Trudeau
will live forever ." Smith , ever forthright and voluble , unhesitatingly
shouts out , " Trudeau
these considerations , it
seems reasonable to say that Smith all along believed that Trudeau is
mortal , notwithstanding the implicit status of the belief . The austere
solution looks unacceptable.
The trick is to find a principled way of dividing , among the sentences Smith will assent to , those that are really tacit beliefs " with real
store), but it also brings additional problems . Are all the consequences of all of Smith 's beliefs things he believes? Alas, no, since no
one
will
assent
to
every
such
consequence
, and
some ,
such
as
to . Nor
will
it do to take
consequences
of Smith
's
of us- irrationally , one might say- can fail on occasion to see the
" obvious " consequences of our beliefs (Nisbett and Ross 1980) . On
the other side of the coin, even some of the far-flung consequences
Smith does assent to are not plausible candidates as tacit beliefs if
they were assented to on the spot, so to speak, rather than having
been cranked out by the extrapolator -deducer working behind the
scenes (see especially Lycan (forthcoming )).
Moreover , when we look closely at how the extrapolator -deducer
mechanism might be expected to operate, there are intractable problems . How
does
it " know
" what
beliefs
in
the
store
are
relevant
to
word
will
not
blow
up
in the
next
ten
seconds
. But
that
is not
be
deducible
from
my
other
beliefs
. Like
what
? That
com -
puters never blow up ? (I don ' t know if I believe that .) That North Star
computers never blow up ? That computers only blow up if . . . if
what ? I really don 't have the faintest idea. (See also Dreyfus 1979,
1981.)
beliefs
. It may
be that
on some
occasions
sentences
are stored
, as
for example when the exact words of Macbeth's death speech are
burned
ra ther
the
rule .
what
a female
moose
looks
like
if
that
information
remains
untapped during rutting season. Notice too that the point holds
whether the information is innate or acquired . So much , I think , is
obvious . What is not obvious is how an information -processing sys-
tem knows which of the many things in its store is relevant to the
problem at hand . If everything in the store is accessed, then the
system will be swamped and unable to make a speedy decision (starving to death while it continues to canvass information ), but if relevant
information is unavailable , the organism ' s environment will make it
Functionalist
Psychology 393
pay . As noted earlier, something akin to this question troubled Peirce
profoundly , and the form it took for him was how humans happen to
light upon relevant hypotheses to explain an event.
At first encounter the problem of relevant access might seem
merely a philosopher 's curiosity , but it turns out to be much more .
This becomes most clearly visible in the context of artificial intelli gence (AI ), where the problem is to program a robot that can interact
with the world using a knowledge store. Consider the following example, adapted from Dennett 1984a. The robot 's task is to survive ,
and it is told that the spare battery it needs is in a room where a time
bomb is set to explode soon. Suppose that the battery is on a wagon ,
and, having entered the room , the robot pulls the wagon out of the
room . The bomb explodes and destroys both the battery and the
robot , because the bomb , it turns out , was also on the wagon . Appar ently the robot knew that but failed to see an obvious consequence of
its behavior , so its program must be improved to allow for that .
Having modified the program , we try again. This time the robot
goes into the room and, after pulling out the wagon , begins to crank
out the consequences of its behavior , and the consequences of the
consequences, including such things as " Some heat will be generated
by the action of the wheels on the floor " and " Pulling out the wagon
will not change the color of the walls ," and the bomb explodes while
it is still cranking .
The defect, it seems, is that the robot needs to distinguish between
relevant and irrelevant consequences, and what it needs is a mechanism for lighting on the relevant ones. Our parents would say to us,
l'Use your common sense: the color of the walls does not matter , but
the bomb's still being on the wagon does." But how do we translate
our parents' advice into an instruction the robot can follow ? Should it
have the following instruction in its store: if you are trying to get your
battery away from a bomb, make sure they are not both on the same
wagon , and all else is irrelevant ? This is both too specific and too
general. Many other things could indeed be relevant , such as: if there
is water nearby, don 't put the battery in that ; if there is apple juice
nearby, don't put the battery in that; don't put the bomb in a fire ;
don 't let yourself be near the bomb; if the bomb can be switched off ,
do that ; and so on and drearily on . Moreover , it seems too specific if it
is to mimic human knowledge , since we do not have explicit instruc tions for what to do when finding bombs and babies on wagons, yet
we would typically behave successfully.
Suppose, on the other hand , we give the robot the more general
instruction , " If you find a hazard situated close to your battery , separate them by a good distance." More will still be needed, however -
Functionalist
Psychology 395
mechanisms that subserve that behavior (Kandel 1976, Horn 1983,
Nottebohm 1981) can thus put a squeeze on the questions.
The co-evolution of ethology and neuroscience is already well
underway , and the two have coalescedat some points in what goes by
the name of neuroethology
. (See for example Horn 1983, Nottebohm
1981, Hoyle 1984.) The complaint that the simpler nervous systems
do not display intelligence is premature : we will not kno\,\r what intel ligence is until we have a solution to the problem of knowledge use.
In any event, if there are grades of intelligence , understanding the
lower grades may be essential to understanding the higher grades.l0
(See also chapter 10.)
The problems of tacit knowledge and knowledge access are not
unrelated , and together with the infralinguistic problem they make
an impressive casefor assessingthe sentential paradigm as unsound .
The sententialist is of course right about this much : if cognition is not
in generalthe manipulation of sentences, then we don 't know what is
going on . The antecedent of this conditional seemsto me inescapable,
and since my extrapolator -deducer is in good order , I conclude that
indeed we do not know what is going on . Not , however , that I see
that as cause for despair; it is what we would expect for a science in
statu nascendi. Compare the lament of the vitalist in the nineteenth
century : if there is no such thing as vital spirit , then we do not have
any idea what makes living things alive and different from nonliving
things . Biochemistry eventually changed all that .
Eliminative Materialism
Once folk psychology is held at arm's length and evaluated for
theoretical strength in the way that any theory is evaluated, the more
folkishly inept , soft, and narrow it seemsto be. The crucial first step is
to see that it is a theory , notwithstanding its observational application
to oneself and others, and notwithstanding its overpowering obviousness and familiarity . Insofar as it is a theory , it is an empirical ,
not an a priori , question how good a theory in fact it is, and its
" obviousness" will not protect it from revision or replacement if it is
flawed .
Like other folk theories, folk psychology is not without virtue , and
within certain circumscribed domains it has considerable predictive
success. It has at least been good enough to survive until now . On the
other hand , it would be astonishing if folk psychology , alone among
folk theories, was essentially correct. The mind -brain is exceedingly
complex, and it seems unlikely that primitive folk would have lit
upon the correct theoretical framework to explain its nature where
psychology is misconceived in a number of dimensions . The sentential paradigm seems increasingly problematic , and interest in alternatives is on the rise. In such an environment , new theories, looking at
first damp and frail , will hatch and struggle to fly . (See chapter 10.)
By " eliminative materialism " I mean the view that holds
1. that folk psychology is a theory ;
2. that it is a theory whose inadequacies entail that it must eventually be substantially
" eliminative " ); and
revised
or replaced
outright
(hence
1979 .)
eliminative
materialism
sees
it , the
sentential
attitudes
do
not
of cognitive
activity . As
Hooker reminds us, " Language will surely be seen as a surface abstraction of much richer , more generalized processes in the cortex , a
convenient condensation fed to the tongue and hand for social pur poses " (1975:217) . Sentences are linguistic entities , and language is
recall that humpbacked whales communicate by means of " melodies." Imagine that whales have a theory of " melodic attitudes ," in
virtue of which they explain and predict one another's behavior ,
where having a certain melodic attitude is understood
on the model
to a melody
is conceived
to be essentially
Functionalist Psychology
397
. " However
believe
what
, if there
are no
he says . Or
beliefs
, then
if he believes
what
the
eliminativist
he says , then
can there
theoretical
as believing
framework
we are bound
to describe
the eliminativist
framework
framework
the
eliminativist
is available
. If the
wishes
eliminativist
to criticize
is correct
and
no alternative
in his
criticisms
and if the old framework is revised and replaced, then by using the
new vocabulary the eliminativist 's criticisms could be restated with
greater sophistication
(For example, the new eliminativist might declare, " I gronkify beliefs," where gronkification is a neuropsychological state defined
within the mature new theory .) It would be foolish to suppose folk
psychology must be true because at this stage of science to criticize it
implies using it . All this shows is that folk psychology
is the only
Functionalist Psychology
399
Conclusions
The goal of this chapter has been to defuse the antireductionist arguments of the functionalists . Mental states may be functional states,
but this does not imply that the specification of their functional profile
based in folk psychology is correct, either in general or in detail .11Nor
does it imply that psychology cannot be reduced to neuroscience. The
claims for the autonomy of psychology are therefore misbegotten .
Psychology stands to profit from a co-evolutionary development with
ethology and neuroscience, partly becausethere are common areas of
research, and partly because the sentential paradigm appears to have
fundamental shortcomings . New theories about the nature of infor mation processing and about the nature of information -bearing structures are badly needed. In the next chapter I wish to discuss some
attempts to ascend beyond the level of the single cell- attempts to
theorize about representations and computations in systems of
neurons .
SelectedReadings
Churchland, Paul M. (1981). Eliminativematerialismand the propositionalattitudes.
Journalof Philosophy
78:67- 90.
Churchland, Paul M. (1984). Matterandconsciousness
: A contemporary
introductionto the
philosophy
of mind. Cambridge, Mass.: MIT Press.
Dennett, Daniel C. (1975
). Brain writing and mind reading. In Language
, mind and
knowledge
, ed. K. Gunderson. (MinnesotaStudiesin the Philosophyof Science7.)
Minneapolis: University of MinnesotaPress. (Reprintedin D. C. Dennett (1978).
Brainstorms
. Cambridge, Mass.: MIT Press.)
Dennett, Daniel C. (1984). Cognitive wheels: The frame problem of artificial intelligence. In Minds, machines
, andevolution
, ed. C. Hookway, 129- 151. Cambridge:
CambridgeUniversity Press.
400
Chapter 10
Theories of Brain Function
What we require now are approaches that can unite basic neurobiology and
behavioral sciences into a single operational framework .
Dominick Purpura , 1975
1(),1
Introduction
An enormous amount is known about the structure of nervous systems . What is not understood is how nervous systems function so
that the animal sees or intercepts its prey , remembers where it cached
nuts , and so forth . We are beginning to understand the behavior of
an individual neuron - its membrane properties , the spiking proper ties of its axon , the synaptic phenomenology , its patterns of connec tivity , the transport of intracellular materials , its metabolism , and
even something of its embryological migration and development . To
be sure , many unanswered questions remain , and major break throughs are yet to be made , but the neuron is not a smooth -walled
mystery . Neuroscientists have considerable confidence at least about
what sort of research would lead to answers to outstanding ques tions , and they have a general picture of what , more or less, the
answers will look like . On the other hand , the state of theory of how
ensemblesof neurons result in an owl ' s being able to intercept a zigzag ging mouse , for example , or how any creature visually recognizes a
given object , is markedly different . Here there is no widely accepted
theoretical framework , nor even a well -defined conception of what a
theory to explain such things as sensorimotor control or perception or
memory should look like .
Theorizing about brain functions is often considered slightly disrep utable and anyhow a waste of time - perhaps even " philosophical ."
A neuroscientist randomly plucked out of the crowd at the Society for
Neuroscience meetings and asked about the role of theories in the
discipline will likely answer with one or all of the following : (1) " The
time for theories has not yet arrived , since not enough is known
404
A Neurophilosophical
Perspective
about the structural detail ," (2) " What is available by way of theory is
too abstract, is untestable, and is anyhow irrelevant to experimental
neuroscience ," (3) " You cannot get a grant for that sort of monkey -
then spending yet more time and effort in conducting the experiments . If the theory turns out to be a flop , and in the absence of a
mature paradigm it well may , then the investment may be a career
disaster . 50 the decision to adopt a policy that says " Leave theorizing
. This
is not
an idle
worry
in the
current
state
of neuroscience
searcher has mastered a certain technique , and there are always more
measurements he can make. The technique comes to structure and
govern
the
research
program
rather
than
the
other
way
around
. The
justification offered for such research is the " maybe-mightbe " two step: " If . . . then maybe . . . , and then my results might be
important ."
whose
relevance
is God
knows
what
. The
idea
that
all results
are
405
at least several unblushing instances of the " maybe-mightbe " two step .
In a general sense the best experiments
and
rich
the
available
theoretical
framework
, the greater
the
there is a reason for doing one experiment rather than another , there
must be some governing hypothesis or other in virtue of which the
experimental question is thought to be a good question, and some
conception of why the experiment is worth the very considerable
trouble
. There
significant
must
, that
is , be
some
sense
of how
the
results
are
for the larger picture of how the brain works (Kuhn 1962).
when they are known , there remains the problem of accounting for
how
the ensemble
works . And
the
function
of the
ensemble
cannot
be
406
A Neurophilosophical
Perspective
the participating
neurons
since ,
. Whatever
components
it
is that
do , nor will
ensembles
do , it
it be a summation
will
not
of what
look
like
what
components
do .
it may well be
business
is a " relevant
structural
detail
" may
in fact
be recog -
to explain
ensemble function . For example, unless you think that DNA is hereditary material , you will not think the organization
of nucleotides
is
of the need
of whole
brain
function
The cardinal background principle for the theorist is that there are
no
homunculi
. There
inner television
is no
little
person
in
the
brain
who
screen , " hears " an inner voice , " reads " the topo -
obvious
, and
even
a brief
immersion
in
the
neurosciences
should proof one against the seductiveness of homuncular hypoth eses. Surprisingly , however , homunculi , or at least the odor of
homunculi , drift into one ' s thinking
rassing frequency .
Part of the explanation
presence of homuncular
preconceptions is that folk psychology still provides the basic theoretical framework within which we think about complex behavior . Unless warned off , it insinuates itself into our thinking about brain
function as well . In a relaxed mood , we still understand perceiving ,
407
thinking , control , and so forth , on the model of a self- a clever selfthat does the perceiving and thinking and controlling . It takes effort
to remember that the cleverness of a brain is explained not by the
cleverness of a self but by the functioning of the neuronal machine
that is the brain . (See also Crick 1979.)
Crudely , what we have to do is explain the cleverness not in terms
of an equally clever homunculus , and so on in infinite regress, but in
terms of suitably orchestrated throngs of stupid things (Dennett
1978a, 1978b). In one's own case, of course, it seems quite shocking
that one's cleverness should be the outcome of well -orchestrated
stupidity . The sobering reminder here is that so far as neuronal organization is concerned, there appears to be no rationale for giving a
system conscious access to all- or everl very many- of the brain 's
states and processes.
10.2 In Search
of Theory
What is available by way of theory ? Are there theories that have real
explanatory power , are testable, and begin to make sense of how the
molar effects result from the known neuronal structure ? Less demandingly , are there theoretical approaches that look as though they
will lead to fully fledged theories?
The fast answer is that a lot of very creative and intelligent work is
going on in a number of places, but it is uneven, and it is difficult to
determine how seminal most of it is. I began scouting the theoretical
landscape with neither a clear conception of what I was looking for
nor much confidence that I should recognize it if I found it . Most
generally , I was trying to see if anywhere there was a kind of " Galilean combination " : the right sort of simplification , unification , and
above all, mathematization
- not necessarily a fully developed theory ,
but something whose explanatory beginnings promised the possibility of real theoretical growth .
In coming to grips with the problems of getting a theory of brain
function , I had to learn a number of general lessons. First, there are
things that are advertised as theories but are really metaphors in
search of a genuine theoretical articulation . One well -known example
is the suggestion first floated by Van Heerden (1963) that the brain 's
information storage is holographic . (Seealso Pribram 1969.) Now the
brain is like a hologram inasmuch as information appears to be distrib uted over collections of neurons . However , beyond that , the holographic idea did not really manage to explain storage and retrieval
phenomena . Although significant effort went into developing the
analogy (see, for example, Longuet -Higgins 1968), it did not flower
408
A Neurophilosophical
Perspective
409
is offered as the theory itself , but evidently such a list is not per se a
theory of what processes intervene between input and output . A list
may include items such as that the brain is in some sense selforganizing , that it is a massively parallel system, and that functions
are not discretely localizable but in some sense distributed . But a list
of this sort does not add up to a theory, though the items are relevant
considerations to be stirred into the pot . Like the prohibition against
homunculi , they might be construed as constraints that any serious
theory will ultimately have to honor . Or in more old-fashioned language, they might be called " prolegomena for future theorizing ."
(See also below .)
Fourth , as Crick has said, it is important to know what problems to
try to solve first , and what problems to leave aside as solvable later .
Becauseone is ever on the brink of being thrown into a panic by the
complexity of the nervous system, it is necessary at some point to put
it all at arm' s length and ask: What answers would make a whole lot
of other cards fall ? What are the fundamental things a nervous system
must do? This of course will be a guess, but an educated guess, not a
blind one. The hope of any theorist is that if the basic principles
governing how nervous systems operate are discovTered
, then other
operations can be understood as evolution 's articulation and
refinement of these basic principles . Simplifications , idealizations ,
and approximations , therefore , are unavoidable as part of the first
stage of getting a theory off the ground , and the trick is to find the
simplification that is the Rosetta stone, so to speak, for the rest. In
physics, chemistry , genetics, and geology, simplifying models have
permitted a clarity of analysis that lays the foundation for coping with
the tumultuous complexity that exists. Accordingly , Ramon y Cajal's
warning against II. . . the invincible attraction of theories which simplify and unify seductively " should not be taken too much to heart . If
a theory is on the right track, then the initial simplifications will grow
into more comprehensive articulations ; otherwise , it will shrivel and
die .
The guiding question in the search for theory is this : What sort of
organization in neuron -like structures could produce the output in
question , given the input ? Different choices will be made concerning
which output and input to focus on. For example, one might select
motor control , visual perception , stereopsis, memory , or learning
about spatial relations as the place to go in . What is appealing about
yTisualperception is that \-\ITeknow a great deal about the psychology
of perception and about the physiology! of the retina, the lateral
geniculate nuclei , and the visually responsive areas of the cortex.
Wha t we do not understand , among other things , is how to charac-
410
A Neurophilosophical
Perspective
411
widely concerning what has promise and where the gold is. Generally speaking, theoretical approaches originating with neuroscientists
are decried by those in the computer science business as " computationally naive" ; on the other side of the coin, neurobiologists usually
deplore the " neurobiological naivete" of those whose theories origi nate in computer science laboratories. So long as there is no theoretical approach known to do for neuroscience what Newton did for
physics, we are all naive. Inevitably , there is a tendency to see one's
own simplifications as " allowable provisionally " and someone else's
as a fatal flaw . To one convinced of the gold in his own bailiwick ,
other theoretical diggings may seem crackpot . Additionally , if a
theory has quite grand ambitions , it stands to be derided as " pie-in the-sky" ; if , on the other hand , a theory is narrow in scope and highly
specific, it risks being labeled " uninteresting ."
My approach here will be to present three quite different theoretical
examples with a view to showing what virtues they have and why
they are interesting . Each in its way is highly incomplete ; of course
each makes simplifications and waves its hands in many important
places. Nevertheless, by looking at these approaches sympathet ically , while remaining sensitive to their limitations , we may be able
to see whether the central motivating ideas are powerful and useful
and, most importantly , whether they are experimentally provocative .
My strategy can be defended quite simply : if one adopts a sympathetic stance, one has a chance of learning something , but if one
adopts a carping stance, one learns little and eventually sinks into
despair .
Regardless of whether any of the three examples has succeeded in
making a Grand Theoretical Breakthrough , each illustrates some im portant aspect of the problem of theory in neuroscience: for example,
what a nonsentential account of representations might look like , how
a massively parallel system might succeed in sensorimotor control ,
pattern recognition , or learning , how one might ascend beyond the
level of the single cell to address the nature of cell assemblies, how coevolutionary exchange between high -level and lower -level hypoth eses can be productive . They all try to invent and perfect new
concepts suitable to nervous system function , and they all have their
sights set on explaining macro phenomena in terms of micro phenomena . Being selective means that I necessarily leave out much
important work , but given the limitations of space, that is something I
can only regret, not rectify .
Two of the examples originate from within an essentially neurobiological framework . The first focuses on the fundamental problem of sensorimotor control and offers a general framework for
412 A Neurophilosophical
Perspective
understanding
The
the
authors
Llinas
of
and
aticity
their
able
length
owing
to
seeks
The
sandwich
between
of
or
tionist
operations
appropriate
in
there
possibility
ing
the
co
- evolution
useful
to
place
within
the
of
The
one
thing
known
start
of
limited
to
its
The
in
then
of
of
plied
by
climbing
two
and
to
the
is
phase
to
generate
to
this
use
the
in
net
approach
is
artificial
neurobiology
where
( of
concern
and
it
the
will
network
to
concern
,
be
theory
first
fumblings
coordinate
theory
be
the
micro
transforma
cerebellum
detail
almost
( five
of
cerebellum
two
each
cortex
is
more
cell
the
) ,
,
nature
the
has
the
. The
made
each
Purkinje
neuronal
) ,
and
of
the
mossy
of
of
input
and
has
just
one
it
type
is
and
sup
the
possible
distinct
2 .4 ) .
input
fibers
each
con
( figure
job
anon
organization
fiber
fibers
some
it
manner
systems
the
because
positioned
exclusive
cerebel
been
incoming
one
regimented
which
the
has
plus
and
highly
everything
- organization
properties
climbing
: the
that
approachability
investigable
started
said
the
different
in
fiber
drawn
tensor
led
and
the
can
types
very
study
be
hypotheses
what
spaces
electrophysiological
mossy
of
experimental
cell
,
to
of
it
cerebellar
. This
the
neuron
and
the
fibers
determine
of
neuron
Purkinje
just
been
to
paradigm
structural
discussion
inception
about
characteristic
of
the
has
by
"
and
distinctive
output
cell
connec
standard
constrained
neuroanatomists
dream
of
morphologically
nected
two
unlike
wider
) . Connec
elements
other
lessons
functional
know
For
number
one
the
be
computational
try
- like
but
and
exaggeration
the
to
" neurocognitive
idea
want
is
which
name
strategy
neuron
informed
new
shape
some
would
lum
the
the
of
( POP
systems
to
opening
of
to
the
out
and
philosophical
a
general
only
science
is
of
took
place
With
it
the
the
slowly
figure
contrast
general
context
how
tions
to
by
by
can
within
goes
the
and
discussion
processing
from
In
focused
development
and
examines
Theory
are
the
and
Network
ing
new
consider
specified
\ rly
distributed
computer
at
mechanisms
distributed
.
in
system
it
Crick
Rodolfo
general
discuss
approach
systems
parallel
research
Tensor
phenomena
based
intelligence
tive
nervous
of
is
trying
arrangement
essentially
Because
systems
and
the
shall
narroV
third
parallel
are
and
I
Francis
more
research
of
macro
.3
two
modeling
used
is
. The
these
models
scope
attentional
This
nervous
Pellionisz
with
certain
.
researchers
- like
broad
encompass
intelligence
the
computer
of
for
of
Andras
originating
succinctly
artificial
tionism
10
basis
quite
work
very
to
hypotheses
discussed
are
the
second
neurobiological
architecture
approach
theory
psychological
field
computational
this
to
class
output
population
of
rela
413
in the cerebellum is huge- something on the order of 1010neuronsand there is at least another order of magnitude in synaptic connections . Nonetheless, basic structural knowledge of the cerebellum has
made it possible to construct a schematic wiring diagram that illus trates the pathways and connectivity patterns of the participating
cells (figure 10.1). The first point , then, is that a great deal is under stood at the level of micro -organization .
Exactly what the cerebellum contributes to nervous system function
is not well understood , however . What is known is that it has an
important role in coordinating movement , as well as in moving the
whole body . It is what permits one to smoothly touch one' s nose,
catch an outfield fly , or land a snowball on a passing car. The complexity underlying any of these feats puts high demands on a nervous
system. For example, in catching a fly ball, a baseball player must
estimate the trajectory of the ball and keep his eyes on it while run ning to where it is expected to fall . So he has to run , visually track,
maintain balance, reach to intercept , and finally catch the ball .
Subjects with cerebellar lesions show a decomposition of movement, almost as though the various parts of each movement had to be
thought out one by one. Undershooting and ovTershootingthe target
and moving the limb in the wrong direction are also typical d ysmetric
signs in cerebellar subjects. Cerebellar patients also have difficulties
in checking a fast movement , such as a swing of the arm . There are
commonly problems in gait, showing themselves especially in unsteadiness and large stride . Depending on the area of lesion, there
may also be motor impairment of speech (dysarthria ). Playing
baseball is out of the question .
It is known that the cerebellum is not essential for movement because subjects with a nonfunctioning cerebellum can still make volun tary movements . But evidently it is essential for well -controlled ,
"","'ell-timed , well -spaced movement . Plasticity in the nervous system
does permit some compensation in the event of cerebellar lesions
occurring early in development . Children whose cerebellar hemispheres are damaged early in life may nonetheless develop quite
good motor control , so long as the more medial structures in the
cerebellum (the flocculonodular lobe and the vermis ) are undamaged .
But if these structures are also damaged and the entire cerebellum is
nonfunctional , the child remains ataxic (that is, suffers deficits in
motor coordination ) and d ysmetric .
The evolution in complexity and size of the cerebellum in humans
is at least as striking as that of the cerebrum. Correcting for body size,
humans have a larger cerebellum than, for example, chimpanzees,
whose cerebellum in turn is larger than that of horses or dogs. As one
414
A Neurophilosophical Perspective
Stellate
cell
Molecular
layer
Purkinje+
neuron
Purkinje layer
Goigicell
Granular
layer
Mossy fiber
Electric
synapse
- -
// /
Inferior
olive
Intracerebellar
. ""' ,
I o- r
-\
andvestibular
+ Spinalcord
Out
Reticular
nuclei
Pontine
nuclei
I
,
..I
...
.... ..
" "
<] Excitatorysynapse
~ Inhibitory synapse
- - - - - - - - -- ---- - ~._ ._- -- ------ - Spinal Cord
Figure 10.1
Circuit diagram for the cerebellar cortex. Purkinje cells are excited directly by climbing
fibers and indirectly (via parallel fibers from the granule cells) by the mossy fibers.
Stellate and basket cells, which are excited by parallel fibers, act as inhibitory inter neurons . The Golgi cells act on the granule cells with feedback inhibition (when excited
by parallel fibers) and feedforward inhibition (when excited by climbing and mossy
fiber collaterals). The output of the Purkinje cell is inhibitory upon the cells of the
intracerebellar and vestibular nuclei . (Modified from Chez and Fahn (1981). Ch. 30 of
Principlesof Neural Science
, ed. E. R. Kandel and J. H . Schwartz, pp . 334- 346. Copyright
1981 @ by Elsevier Science Publishing Co., Inc.)
might predict , therefore , human versatility in motor control is remarkable . To mention only a tiny sample, we can swim , pole-vault ,
climb trees, use knives , speak languages, whistle , draw ~ skate, and
play musical instruments . For each of these accomplishments the
nervous coordination of muscles is a stunningly complex affair .
What is the input on which the cerebellum can work its miracles? It
includes massive inputs from the cerebral cortex- the motor strip
and nearly everywhere else- as well as from other brain regions
415
subserving motor function . The cerebellum also receives afferent in pu ts from all types of sensory receptors. Some of the cortical in pu ts
are thought to be grossly specified motor commands for which the
cerebellum provides the finely tuned , detailed commands. (This will
be elaborated below .) In the absence of the cerebellum the motor
commands of the cerebral cortex are conveyed down the spinal cord
without the coordinative tuning of the cerebellum .
Now if we know so much , in a general fashion , about what the
cerebellum does, and if we know so much about the fine-grained
structural facts, we ought to be able to figure out hoUJthe cerebellum
does what it does. We ought , that is, to be able to explain how the
activity of the collections of cells produces coordinated movement .
For anyone who hoped that the theor)' would simply tumble out once
so many details \\rere available, the cerebellum seemed strangely frus trating . Because what remained mysterious was the functional
story- intermediate between the gross functional description and the
wiring diagram - that would explain exactly what role the cerebellum
plays in the administration of motor control . The epistemological situ ation provoked diverse researchers into trying to find a fruitful
theoretical orientation (for example, Braitenberg and Onesto 1961,
Marr 1969, Ito 1970).
The line that Pellionisz and Llinas pursued depended on their determination to take as the starting point the parallel nature of infor mation processing in the brain , and in the cerebellum in particular . If
the cerebellum has a parallel architecture , in the sense that many
channels are simultaneously processing information , then, they argued, it is a fair assumption that the computational processes are
suited to that architecture . To understand what the computational
processes might be, they followed the idea that they needed to know
about the patterns of activity within large arrays of neurons .
A " wiring diagram " of cerebellar neurons is useful in describing in
a highly schematic way the connections between input and output .
Typically , however , the diagram displays one or two schematic
neurons and their connections, whereas in fact these are embedded
in an array of thousands of cells. That is, the massively parallel nature
of the network is, for graphic convenience, suppressed. Such suppression will not matter if the schematic neuron is a faithful representative of every neuron in its array- if , that is, the system is essentially
redundant . On the other hand , if the global connectivity pattern
within the array is itself crucial to how the array processes informa tion , then we pay for the convenience of the suppression, inasmuch
as we mask exactly the detail we need in order to understand the
system.
416
A Neurophilosophical
Perspective
417
gan to think that what the network of cerebellar cells did to its input
could be characterized by means of a tensor - a generalized mathe -
what
the frames
of reference
involved
. The
basic
mathematical
nate system, then a tensor is what effects the mapping or transforma tion
from
one
transformation
vector
to the
other
. Which
tensor
matrix
governs
the
in the connectivity
rela -
in terms of their values as specified in relation to the relevant coordinate axes (figure 10.2). If each neuron in a network of input neurons
specifies an axis of a coordinate system , then the input of an individ -
ual neuron - its spiking frequency - defines a point on the axis, and
the input of the whole array of neurons can then be very neatly given
. (For an introduction
1986 ) .
Given the data on input vectors and output vectors supplied in the
model , Pellionisz observed that from a mathematical standpoint , the
connectivity relations between input and output neurons serve as a
matrix , such that any input vector is transformed into an output
vector . That is, the nature of the regularity in the patterns of activity
of the neuronal arrays represented in the model invited the hy pothesis that the arrays are doing matrix multiplication. In particular ,
the systematic differences in response profiles of sets of Purkinje cells
situated at different locations on the same beam of parallel input
418
A Neurophilosophical Perspective
V = (3, 2)
M
[- 6
2-4
5
3]
1
tern to another
. What
coordinate
by the
representational
job of a given ensemble .
To begin with , there will be the coordinate
system specified
by
visual or olfactory
or vestibular
input arrays and the very different
coordinate
system specified
by motor output
arrays . Suppose , in deed , that the fundamental
computational
problem
of sensorimotor
control is the geometrical
problem of going from one coordinate
sys tem (e .g ., visual ) to another , very different
coordinate
system (e .g .,
motor ) . Then arrays of neurons are interpretable
as executing vector to -vector transformations
because that is what they really are doing the computational
problems
a nervous
system has to solve are
fundamentally
geometrical
problems . The idea seemed
to have
plausibility
not only for the cerebellum
but for wider domains
as
well .l
A tensor is a generalized mathematical
function
for transforming
vectors into other vectors , irrespective
of the differences
in metric and
dimension
of the coordinate
systems . If the basic functional
problem
of sensorimotor
control
is getting
from
coordinate
ensemble
geometric
embodiment
of a tensor .
of the problem
of sensorimotor
420
A Neurophilosophical
Perspective
J
-
'
,
,
-
Figure 10.3
The problem of sensorimotorcoordination. (a) and (b) depict a crab-like robot wi~h
rotatableeyesand an extendablearm. As the eyes triangulatea target by assuming
angles(a,r3), the arm joints must assumeangles(6, 'P) suchthat the tip of the forearm
makescontactwith the target. (Adapted from Paul M. Churchland(forthcoming).)
421
from the straight ahead position (see figure 10.3). It is most convenient to characterize the straight ahead position as 90; hence, all eye
positions can be specified as the angle subtended by the horizontal
axis and the " fovea line ." For example, suppose that the apple is at
(1.2, 10.8) in external space. Then to foveate it , the eyes must rotate:
the left , such that the angle subtended by the horizontal and the
" fovea line " is 65 degrees; the right , such that the relevant angle is
105 degrees. In Roger's visual phase space, therefore, the position of
the apple is given by the ordered pair of position angles for each eye,
namely (65, 105).
Roger' s visual space is a phasespacein which the position of the
apple is represented by the two eye-rotation angles, alpha and beta,
that jointly triangulate it . Any coordinate system specifies a phase
space, and here " position in phase space" simply refers to the global
condition of the physical system being represented. Phase spaces
may differ as a function of the number of coordinate axes (2, 3, 50,
10,000, etc.) and the angles of their axes (at right angles to each other ,
or nonorthogonal ). A hyperspace is a phase space with more than
three dimensions . A phase space may be Euclidean, but it need not
be. It could , for example, be Riemannian, in which case the interior
angles of a triangle inscribed in that space need not sum to 180
degrees. (A caution : the term " phase space" is commonly used in
classical mechanics to denote a specific coordinate space of six dimen sions, three for position and three for momentum . But as I use the
term here, it has the entirely general meaning of " coordinate space"
or " state space." See also Suppe 1977.)
Note that for any position of the apple in external space, there is a
corresponding position of the apple in Roger's visual phase space.
Accordingly , we can say that Roger's visual vector, such as (65, 105),
representsthe position of the apple in the world , since there is a
systematic relation between where the apple is in the world , as described in external coordinates, and " where " in visual phase space it
is, as specified by a pair of eye-angle coordinates (figure 10.4a).
Just as Roger has a 2-D visual space in which the position of the
object is represented, so he has a 2-D motor space in which its arm
position can be represented . But, and this is crucial, these two phase
spaces are very different . How do we characterize Roger's motor
phase space? Again , by specifying the axes appropriate to his motor
equipment . This time , the position of his limb in phase space is given
by the two angles by which it deviates from a standard position .
Thus, let the zero position of the upper arm be flush with the horizon tal axis. Then a position of 45 on the upper -arm axis will represent an
upper -arm position of 45 degrees off the horizontal (figure 10.3). Cor-
422 A Neurophilosophical
Perspective
a
~
sensory
phase space
Figure 10.4
The respective configurations of the crab's sensory and motor systems can each be
represented by an appropriate point in a corresponding phase space: (a), (b). The crab
needs a function from points in sensory phase space to points in motor phase space.
(Adapted from Paul M . Churchland (forthcoming ).)
423
relatively , let the zero position of the forearm be its position when
extended straight out from the upper arm, wherever the upper arm
happens to be positioned . Accordingly , a position of 78.5 on the
forearm axis represents the forearm as rotated 78.5 degrees
coun terclockwise from the line extending out from the upper arm,
whatever
the position
can give the overall position of Roger' s arm in motor phase in terms of
the two angles as (45, 78.5). Moreover , we can specify the position of
the apple by specifying that arm position where the tip of the forearm
just touches the apple . The sensory phase space and the motor phase
space are represented
Now for the action (figure 10.5). The eyes, having detected the
apple, announce its position : " Apple at (55, 85)." Notice that if the
arm were to take as its command , " Go to (55, 85)," then it would
424
A Neurophilosophical
Perspective
425
~
~
-- : >
<
90
W
w
i
:r:
JR
THEY
REACH
FAIL
THE
CRAb
;: GE
FYF
BEYOND
' S
ARM
90
I FFT
~ ~ GE
AREA :
COr -jYEr
CO ' jV
OF
TO
I
180
AN (', I F
Figure 10.6
The coordinate transformation , graphically represented. The grid in (a) represents the
set of points in sensory phase space that correspond to a triangulated object within
reach of the crab's arm . For each such point in (a), its corresponding position in motor
phase space is entered in (b). The entire set of corresponding points in (b) displays the
global transformation of phase-space coordinates effected by the crab's coordinating
function . The heavily scored triangle and rectangle illustrate corresponding positions
in each space. (Adapted from Paul M . Churchland (forthcoming ).)
lengths , a
426
A Neurophilosophical
Perspective
invariant, which the visual system and the motor system both represent, each in its different way , while the coordinate transformer tells
the effector system what it must do to make contact with the in variant . In geometrical terms, the coordinate transformation tells us
how we have to deform one phase space to get at the object in the
other phase space. Tensors are a means whereby the nervous system
can represent the very same thing many times over, despite the dif ferences in coordinate systems in which the thing is represented. In
sum, then , representationsare positionsin phasespaces
, and computations
are coordinatetransformationsbetweenphasespaces
.
Bear in mind , however , that even equipped with a coordinate
transformer to translate sensory locations into the correct motor locations , Roger is radically simplified - so much so that he would not
stand a chance in the real, cutthroat , biological world . Even the
humblest nervous systems are more complex and more sophisticated
than Roger's. To begin with , his is a 2-D world , and ours is a 4-0
space-time world . As soon as we consider the sensorimotor control
problems that must be solved by the brains of real creatures, making
their living in a 4-0 space-time world , the necessity of elaborating on
the simple coordinate transformer of the cartoon story is plain . More over, Roger never moves his whole body , and he never has the problem of maintaining posture and balance. All he does is visually locate
the apple and then touch it . He does not flee or hide from any predator , he does not mate, he does not build a nest or dig a hole, he does
not even chew up and swallow the apple he touches. Consider also
that although Roger has merely two phase spaces- one afferent and
one efferent- this is an unrealistic arrangement for a real nervous
system, which could be expected to have some number of intervening
phase spaces as well . Nevertheless, the crux of the Pellionisz-Llinas
approach is this : the sensorimotor problems faced by more realistic
creatures can be understood as reducing at bottom to the same general type of problem that Roger faces- namely, the problem of making coordinate transformations between different phase spaces. And
the solution found for Roger's cartoon world illustrates the general
nature of the solution evolved by organic brains in the real world .
Roger is simplified in a further dimension . He has neither muscles
nor neurons to make muscles contract . When he moves an arm, that
is really just the computer painting lines across the CRT screen. When
a real crab moves a claw, it does so by virtue of the precise orchestration of muscle contraction by neurons . Let us wallowa bit in the
sensorimotor predicament of a real crab foraging for food . Supposing
it spots an edible chunk of fish , it must move toward it , grasp it with
its claw, and get it into its mouth . It has to contend with six legs;
427
moreover, each leg has three joints , each joint is served by at least
two muscles, and each muscle consists of many muscle cells and is
innervated by a large number of neurons . If the object to be inter cepted is itself moving , the control problem becomes very complex.
But it is approachable by using the same basic mathematical idea used
in solving Roger' s problem .
If we can think of the crab' s arm as specifying a phase space, then
the set of muscles concerned may also be thought of as specifying a
phase space, where the positions of each muscle are represented on a
proprietary axis of that space, and where a vector in that space is
determined by the degree of contraction of the component muscles. A
phase space of yet higher dimensions is specified by the motor
neurons innervating the muscles, where each neuron is given a dimension and its firing frequency will be represented as a point along
that axis. Notice that we can expect there to be systematic relation ships between positions in the skeletal phase space, positions in the
muscle phase space, and positions in the neuronal phase space. The
central idea is quite simple : the limbs move the way they do because
the muscles contract the way they do, where that pattern of effects is
in turn caused by a pattern of neuronal activation of the muscle units .
Animals ' motor systems had to evolve systematic relationships
among the phase spaces of motor neurons, muscle cells, and limbs if
they were going to use neurons to control the movement of muscles
and thereby control the movement of limbs . Any animal whose motor
system lacks such relationships will not be able to move properly , and
its survival time will be brief . From the perspective of tensor network
theory , to look for the functional relations between connected cell
assemblies is to investigate the properties of the relevant phase
spaces- that is, to determine their geometries, and to determine the
transformations that will take us from the representation of some
external invariant in a given space to its representation in a different
phase space. Knowing the geometry of the limb phase space, therefore, will guide us in approaching the motor neuron phase space.
Similar points apply of course to matters on the afferent end . Ro'- er
Jdoes not detect the presence of the target by virtue of photosensitive
neurons . Biological organisms with r~al eyes do . Nevertheless, the
basic principle of representation as position in phase space and computation as coordinate transformation can be invoked . That is, in real
organisms retinal neurons will specify a phase space, vestibular
neurons will specify a phase space, and so forth . If the afferent system is to playa role in the organism 's feeding , fleeing, and so forth ,
then afferent phase spaces will have to be coordinated with efferent
phase spaces. Sensory phase spaces are bound to be different from
428
A Neurophilosophical
Perspective
429
Parallel fibre
" summed" at each axon hillock and a spike train is emitted . Thus, the
output vector has as its components x, y, and z, the three output
frequencies. Vectors (input ) are thus transformed into vectors (output ) via matrix multiplication .
Although the model neuron array is highly schematic, it is fairly
easy to imagine how to embellish it . For example, the dimensionality
of the phase spacescan be increased by adding neurons, redundancy
can be accommodated if needed by " twinning " neuron configurations , the matrix can be made plastic by allowing for mod ifiability of synaptic junctions or for the addition of receptor sites.
Moreover , something resembling this neuronal array does exist in the
cerebellar cortex. Indeed , it was precisely by pondering the regimented organization of parallel fibers and Purkinje cells in the cere-
430
A Neurophilosophical
Perspective
bellar
cortex
that
Pellionisz
and
Llinas
came
to
the
view
that
their
are in the
cerebellum
other
matters
to be factored
in , such
as
the role of incoming climbing fibers (one to each Purkinje cell with
multiple synaptic contacts) and the function of neurons in the cerebellar
nucleus
. -~ u -t Pellionisz
and
Llinas
consider
that
these
can
be
accommodated within the basic framework of phase space representation and vector -to -vector transformation
(Pellionisz and Llinas
1982) . If the cerebellum is executing tensorial transformations , the
next question is this : what is the character of those phase spaces such
that
a vector
from
one
is transformed
into
a vector
of the
other
? The
answer to that depends on the empirical facts about what the motor
cortex and the cerebellum are really up to, but at least a rough answer, based on cliI}ical, physiological , and anatomical data, is already
discernible
Crudely , the plot line is this : the input from the cerebral cortex
specifies in a general way what bit of behavior is called for . For example, suppose the incoming " intention " to my cerebellum is " Touch
that (apple) with my right hand ." The incoming " intention " vector
specifies this position in a sensorimotor coordinate system (touch that
seen/heard object), but it does not specify a curve in the motor space
that says exactly how the goal position is to be achieved such that the
target is intercepted . It is the job of the cerebellum to transform that
intention
vector
into
an
execution
vector
that
will
orchestrate
the
sequence
of muscular
contractions
. It will
have
to coordinate
as the
finger closes in on the target, and then , notice, the finger decelerates.
Often a movement
to inform
the
next
motor
command
, and
the
movement
432
A Neurophilosophical
Perspective
This includes not only skin afferents but also muscle and joint afferents. In experiments on monkeys it has been found that the animal
can make good use of deafferented limbs to reach, grasp, climb, and
walk ; moreover , it can learn new movements (Taub 1976). In a rare
case of deafferentation in a human , caused by a peripheral sensory
neuropathy in which the motor neurons were spared, the patient
remained able to perform many motor skills . For example, without
visual feedback he could easily touch his nose, touch each finger
sequentially with the thumb , and draw (on command) circles,
squares, and figure eights . Evidently in these instances he was executing the sequence of movements without benefit of any feedback
at all . Remarkably, this patient was even able to drive a car with a
gear shift , though when he bought a new car he could not learn to
drive it and had to resort to driving the old car (Marsden , Rothwell ,
and Day 1984). Given the touchy and idiosyncratic nature of clutches,
this is not surprising .
Now the general principle proposed for cerebellar motor coordination is that the incoming . " intention " vector is transformed , via a
tensor, into an execution vector that specifies the detailed sequencing
of muscle cell activity . Failure of the cerebellum to function means
that the " intention " vector rather than the " execution" vector is the
motor command directly transmitted down the spinal cord, and the
result is inadequate muscular coordination and inappropriate timing .
The finger overshoots or undershoots , and the movement lacks grace
and smoothness, not unlike Roger's fumble if his visual space coordinates are directly used to specify his arm position in intercepting the
target. Of course, the cerebellum may be doing other things as well ,
and it is also likely that there is not one massive connectivity matrix
for motor coordination , but rather sets and even hierarchies of matri ces. Nevertheless, the hypothesis invites us to see tensorial transfor mation as a fundamental principle of operation .
In sum, the tensor network hypothesis says that a neuronal network implements its general function as a connectivity matrix to
transform input vectors into output vectors. There are, accordingly ,
two important strands in the hypothesis : the first accommodates the
fact that the coordinate systems of neuronal ensembles will specify
different frames of reference but must be systematically related, and
the second accommodates the parallel nature of neuronal networks ,
by proposing that individual neurons in the array contribute the components to the vectors, while the structure of the connectivity between neuronal arrays determines the tensorial matrix . It is by trying
to do justice to the parallel nature of nervous systems that one comes
433
to
fa
thorn
how
they
sensorimotor
The
idea
the
be
the
assessment
and
that
of
thought
Naturally
merely
that
it
real
is
the
whether
tal
We
order
isz
must
to
10
least
robust
free
"
comes
scale
their
to
theoret
they
own
experi
interpretive
Tork
hy
data
after
must
of
difference
in
simply
unify
results
matters
Quite
testable
all
Llinas
ancestor
it
be
and
essentially
makes
must
is
fertile
what
results
it
theory
Pellionisz
and
this
crude
analysis
of
will
the
be
appears
and
it
ex
to
it
be
must
must
open
explanatory
to
yield
capacity
the
cerebellum
potential
useful
timeless
of
explanatory
to
its
function
at
Theory
is
image
contraction
of
fluid
in
ocular
focus
results
of
more
flatlander
control
by
The
the
Let
us
on
other
therefore
and
but
Pellion
closely
world
and
possible
the
as
produced
leave
return
to
the
reality
it
isolated
not
all
even
the
directions
to
ear
one
isolated
the
from
motion
The
vestibulo
other
gaze
aspects
to
head
creature
can
moves
creature
precisely
and
to
idealization
the
if
muscles
dimension
whereby
compensate
neck
inner
useful
though
elements
the
movement
arrangement
object
many
of
contributes
is
is
Reflex
involving
of
Ocular
contraction
eyeball
VOR
neuronal
an
is
the
in
directions
eyes
it
at
apparatus
hereafter
is
retina
controlling
although
look
Vestibulo
affair
the
muscles
treating
the
complex
on
vestibular
VOR
to
and
rather
slip
the
the
reflex
the
it
Network
control
tinue
in
of
theory
look
it
Tensor
including
begin
. 5
Gaze
better
simple
neuronosphere
all
large
them
net
coordination
his
was
because
it
tensor
the
motor
where
in
approach
domains
Jack
it
gives
determine
the
get
Llinas
Roger
at
engage
seen
to
_ at
conviction
network
already
respect
in
or
to
doors
have
with
explanations
experimen
enthusiastic
apply
experimental
it
surprisingly
are
bound
what
wh
literally
the
tensor
coherent
are
can
not
neuroscientists
single
really
the
despite
predicting
seriously
give
that
beginning
and
taken
possible
the
plaining
Others
it
McCoy
In
and
function
ender
real
McCoy
is
it
dead
is
data
brain
Some
about
they
because
their
in
is
how
and
for
new
uncertain
like
envision
research
potheses
very
"
this
look
to
mental
are
whether
should
begun
still
the
Some
and
have
achieve
provides
about
varies
from
approach
is
comes
really
questions
neuroscience
it
experimentally
ical
to
approach
which
answered
within
school
transformations
network
within
addressed
suspicious
tensorial
tensor
framework
can
of
use
that
theoretical
of
could
control
is
for
in
an
the
owl
any
movement
con
of
its
Rotation
of
434
A Neurophilosophical
Perspective
the head. As the head rotates to the right , the eyes track the object by
rotating to the left . Here is a simple way to show yourself how clever
the VOR is. First, stretch a hand in front of your eyes and, while
holding your head steady, wave the hand back and forth quite
quickly . Visually track the hand , and try to keep a steady, clear visual
image. What you will get instead is a smeary image. For the contrast,
keep the hand steady, and move your head back and forth at a good
clip . Now the image of the hand is not smeary, as before, but clear
and quite steady. That effect is owed to the VOR .
How does the VOR work ? First we need to know what neuronal
structures are involved . Principal elements are as follows : First, there
are the semicircular canals of the vestibular apparatus in the ear, three
canals on each side, one in each of the three planes. Although commonly thought to be precisely at right angles to each other , the canals
in some species deviate considerably from the presumed ideal . Second , each eyeball has six extraocular muscles (muscles attached to the
eyeball exterior ) for rotating the eyeball in its socket. Basically, the
VOR system circuit is at least a three-stage affair connecting vestibular receptors, via three stages of neuronal links , to the eyeball muscles. These stages consist of (1) the neurons responding to input
signals from vestibular receptor cells, which synapse on (2) neurons
in the vestibular nuclei (called secondary vestibular neurons), which
synapse on (3) oculomotor neurons that innervate the muscles (figure
3.9). Notice that although only one neuron is sketched in to depict
each stage, in reality of course there is a large array of neurons at each
stage.
The problem for the system to solve is how much each muscle unit
should contract in order that the eyeball move to compensate for the
movement of the head. If we conceive the problem in terms of the
tensor approach, it takes this form : Assume the system wants to keep
a particular object in view while the head turns . Then the system
needs to convert a new " head position " vector into a new " muscleposition " vector . The input space will have three dimensions , one for
each canal, and the relevant point along the axis for each canal is
determined by the angle from the " initial " position (figure 10.9). As
the head deviates from the initial position , we can describe its movement as a sequence of points in this vestibular phase space. For that
sequence of points , we must find the " compensating" sequence of
points in the muscle phase space. If the vestibular coordinates are
directly used for specifying where the muscles should go, then the
eyeball will end up in the wrong place, much as Roger's arm ends up
in the wrong place if visual coordinates are not transformed into the
coordinates of his motor space. For simplicity , the diagram assumes
435
Figure 10.9
The vestibular semicircularcanals (A, anterior; P, posterior; H, horizontal) can be
characterizedby their rotationalaxes. For the motor system, the rotationalaxesof the
eyecorrespondto the pull of extraocularmuscles(LR, lateralrectus; MR, medialrectus;
SR, superiorrectus; IR, inferior rectus; SO, superioroblique; 10, inferior oblique). The
rotationalaxesof the vestibularsemicircularcanalsand the extraocularmusclesconstitute two built-in framesof referencefor the CNSto measurea headmovementand to
executea compensatorygaze-shift. (From Pellionisz(1985). In AdaptiveMechanisms
in
GazeControl, ed. A. Berthozand G. Melvill Jones. Copyright @1985by ElsevierScience
PublishingCo., Inc.)
that the head is moving purely horizontally , so that the input maximally registers yaw , as opposed to pitch and roll .
There are six extraocular muscles, so the muscle phase space is sixdimensional , and the muscle vector will have as its components the
points on the axes for each muscle. Where any muscle is on its axis
will be a function of its degree of contraction from a standard position , say when the eyeball is positioned directly ahead. Experimental
data are available expressing the relation between muscle contraction
A Neurophilosophical Perspective
436
PHYSICAL
ROT
ATIONAL
AXES
GEOMETRY
VESTIBULAR
OCULAR
' y'
I
r\11R
1
ISOJ
IR
I SR
10
Z
~I ~
I
LA
LR
MR
....
.
SA
- .
IR
so
--
10
-
0 .000
0 .674
- 0 .674
- 0 ,927
0 . 155
0 . 155
0 .374
0 .722
0 .722
data
from
Blanks
et al . ( 1975 )
Figure 10.10
The sensory and motor systems of coordinates of the VOR, intrinsic to CNS function ,
as defined by the vestibular matrix and eye muscle matrix . The directions in threedimensional XYZ space of the unit rotational axes, belonging to individual eye muscle
contractions , are shown on the left ; that is, this illustration shows how the eyeball
moves in 3-space relative to the activity of a given muscle. The excitatory activation axes of the combined semicircular canals of the two vestibuli are shown on the right .
These two frames of reference therefore delimit the phase spaces within which the
nervous system must function such that gaze control is achieved. To facilitate visual
perception of the three-dimensional directions of the axes, their orthogonal projection
to the XYZ plane is also indicated . The numerical values are based on anatomical data.
(From Pellionisz (1985). In Adaptive Mechanismsin GazeControl, ed. A . Berthoz and G.
Melvill Jones. Copyright @ 1985 by Elsevier Science Publishing Co., Inc .)
437
transforms a premotor vector into a motor vector that tells the muscles what the position in muscle phase space should be- in other
words , how much each muscle should contract (figure 10.11). Since
we can figure out what the positions of the eyeball should be given
the position of the head, we can determine the tensorial transforma tions needed. Then we can work backward and figure out what the
participating neurons at each stage should be doing . This in turn can
be tested by seeing whether the neurons really do behave as the
hypothesis says they should . As more is discovered about the
neuronal basis, the basic hypothesis may be corrected and elaborated,
and thus theory and experimental research co-evolve .
To give an example of a testable hypothesis emerging out of the
theoretical considerations : Pellionisz reasoned that the vestibular apparatus should have a preferred position , called the eigenposition , in
which the output vector is different merely in magnitude from the
input vector, and thus the tensorial transformation is maximally simple . Mathematically it can be shown that there is indeed a set of such
positions , though it differs from species to species. In humans the
head' s " best" position ought to be tilted slightly upward at a pitch of
21 degrees; in rabbits, at 24 degrees. This in fact appears to be so.
The tensor network strategy then exploits this idea: there should be
a systematic relation between the vestibular phase space and the
neurons that , by their pattern of firing frequencies, represent the
position of the head. That is, there should be a systematic relation
between the phase space of the vestibular canals and the phase space
of the vestibular neurons, and similarly for the eyeball muscles and the
oculomotor neurons . If we can generate hypotheses about the phase
spaces of the sensory and motor systems, this allo\l\lTsus to generate
fine-grained hypotheses about the neuronal implementation and to
test the hypotheses against the facts of neuronal responses. This is
the framework for experimental investigation .
A further test of the hypothesis is how well it fares in computer
simulation . The answer is that the Pellionisz computer model (1985)
of gaze control involving the VOR and neck muscle coordination is
certainly impressive . It uses a realistic basis for specifying components of vestibular vectors and oculomotor vectors. One virtue of a
computer model is that it permits us to test the theory by asking the
model whether the neuronal array could be executing tensorial transformations on its input , given the experimentally constrained
configuration of neurons and their electrophysiological properties in
the model . If the answer in a highly constrained computer model is
no, then the answer forthcoming from the brain itself is not likely
to be yes. The answer from the Pellionisz-Llinas computer model
439
appears to be that the VOR neuronal array could indeed be transforming covariant vectors into contravariant vectors- that the connectivity and electrophysiology could support such a function .
As indicated at the outset, matters are much more complicated than
the three-stage characterization of the VOR implies . No circuit is as
isolated from other fields of .neuronal activity as the schematic description of a reflex would permit us to suppose. As Sherrington
observed (1906), " a simple reflex is probably a purely abstract conception , because all parts of the nervous system are connected together
and no part of it is probably ever capable of reaction without affecting
and being affected by various other parts . . . ."
First of all, there are connections to the vestibular nuclei from the
cerebellum, which figure in the saccadicmovement of the eyes. Moreover, since not all objects are tracked, there must be factors in virtue
of which the tracking response is differentially engaged, according to
interest , motivation , and such. The neck muscles are a crucial part of
the story since the head moves, and there is feedback from them .
And from behavioral output , we know there must be connections to
visual perception . Additionally , the VOR is in some measure plastic,
since it can adjust to reversing prisms on the eye (Gonshor and Jones
1976).
These complexities can be addressed within the framework of tensor network theory , and with the basic schema in place it is possible
to envision how to begin to factor in additional features. In this spirit ,
Pellionisz has sketched a schematic frog nervous system that gives a
rough picture of how tensor network theory means to encompass
representation and computation from the initial sensory input to the
final motor output (figures 10.12, 10.13). For example, in this sketch
reception vectors from the visual and auditory receptors are inte grated in the superior colliculus by a tensorial transformation . The
idea is that different modalities can be understood as different axes in
a phase space. A position in the multimodal phase space will be
determined by the components of the various axes, and what is perceived is hence a unified , objective target, as opposed to the-objectas-seen or the-object-as-heard .
Further elements in the sketch envision the sensory cortex determining whether the target should be snapped at, the vestibulocerebellum determining how posture would need to be corrected in that event, the motor cortex relaying a vector specifying
what configuration the body should go into relative to the target, and
the cerebellum transforming that vector into a vector that specifies in
detail how the body (hindlimbs , neck, forelimbs ) should achieve that
configuration , given its starting position , in order to converge on the
440
A Neurophilosophical
COVARIANT
Perspective
JIISENS
~RY
RECEPTION
#/
optic - auditory
target
reception
to
e-f
~ -c.-- fore
limb
Figure 10.12
Symbolicdepiction of natural framesof referencein which multisensory-multimotor
transformationsareimplementedin neuronalnetworksof the frog CNS, perhapsin the
tectum. In this examplethe CNS integratesinformation from severalsensorymodalities and transforms the coordinatedsensoryoutput of the colliculus into different
optional motor responses
. The coordinationproblemis complex, becausein snapping
at a fly, the frog must use coordinatedinput from vision, audition, and the vestibular
system, and snappingmay involve a "whole body saccade
" to a target, usinghindlimb
muscles, and stabilizationof the headby neckand forelimb muscles. (From Pellionisz
1983b
.)
442 A Neurophilosophical
Perspective
The hypothesis is best understood as a special case of the more
general theoretical approach advocated by Pellionisz and Llinas . In
more detail the idea is this :
If we suppose that our crab [Roger; section 10.4] contains internal
representations of both its sensory phase space and its motor
phase space, then the following arrangement will effect the desired transformation . Let its sensory phase space be represented
by a physical grid of signal-carrying fibres, a grid that is metri cally deformed in real space in just the way displayed in figure
[10.14]. Let its motor phase space be represented by a second
grid of fibres, in undeformed orthogonal array . Place the first
grid over the second, and let them be connected by a large num ber of short vertical fibres, extending downwards from each vertex in the sensory grid to the nearest vertex in the underlying
motor grid , as in figure [10.14] .
The system functions as follows . Depending on where each
eye is positioned , a signal is sent down a particular fibre in the
appropriate bundle , a fibre which extends into the sensory grid
from the appropriate radial point . Joint eye position is thus reliably registered by a joint stimulation at a unique coordinate
in tersection .
In the lower grid , a stimulation at any coordinate intersection
is conveyed outwards along the relevant pair of motor fibres,
each of which induces its target limb joint to assume the appropriate angle. We need now suppose only that the vertical connections between the upper and lower grids all function as
" and-gates" or " threshold switches" , so that a stimulatory signal
is sent down a vertical connection to the motor grid exactly if and
exactly where the relevant sensory intersection point is simulta neously stimulated by both of its intersecting fibres . Such a system will " compute " the desired coordinate transformations to a
degree of accuracy limited only by the grain of the two grids , and
by the density of their vertical connections. Because the upper
map is metrically deformedin the manner described, the points in
sensory space are brought into appropriate register with the
points in motor space. As the eyes fixate, so fixates the arm: it
reaches out to precisely the point triangulated by the crab's eyes.
(Paul M . Churchland (forthcoming
For reasons reflecting the overtly geometrical nature of the idea,
Churchland has called this the phase-spacesandwich hypothesis . By
spatially organizing themselves into maps, and by layering so that the
maps are in suitable register, neurons might , with utmost simplicity ,
443
COORDINATE
TRANSFORMATION
BY
CONTIGUOUS
TOPOGRAPHIC
MAPS
EYES
\
\
\
\ ~-- ----""""
,II
'n
\.D
SrNSORY
TOPU ~ HAPHIC
( /.\ fTKICA
MA ~
, LL
"
n F ~ 0 R \ 1~ [ ~ )
. . '
."
~~
/
ARM
"/
"
MOTOR
- - - - - TOPOGRAPH
I C MAP
../
(e,cp) = f (~,p)
Figure 10.14
Thecoordinatetransformation, physicallyimplemented. Jointeyepositionis registered
in the upper map by a simultaneousstimulation at a unique intersection. Acting as
and-gates, the vertical connectionsconvey a signal down to the motor map from
exactly that position. From there, activity is conductedout both of the intersecting
motor fibers, with the result that eacharm joint assumesan angleappropriateto the
fiber containingthe signal. (Adapted from PaulM. Churchland(forthcoming).)
444
A Neurophilosophical Perspective
Evidence that some parts of the nervous system may in fact conform in some measure to the phase space sandwich configuration can
be found in studies on the superior colliculus (also known as the optic
tectum, especially in lower animals). High on the midbrain are two
pairs of lumps , the uppermost pair being the superior colliculus (see
section 3.2). Known to be important in foveating a target, the superior
colliculus has a laminar architecture, in which the topmost lamina
receives input from the retina and is organized to map the retina
topographically . The bottom layer sends motor output to the extraocular muscles and is topographically mapped so that stimulation of a
visual cell will cause the eyes to move so as to foveate that part of the
visual field represented by the stimulated cell in the overlying visual
layer . It looks as though the retinotopic map and the extraocularmuscle map are pretty much in register, with the result that vertical
connections (direct or stepwise) from the retinotopic map can excite
the motor neurons to move the eyeball so that it foveates on that part
of the visual field wherein the target was first detected (figure 10.15).
This sketch is an oversimplification , of course. For one thing , in mammals the superior colliculus is a tightly integrated part of a larger
complex system that includes inputs from the visual cortex and the
frontal eye fields . Nevertheless, the suggestion is that the phase
space sandwich may constitute the fundamental organizing plan in
the superior colliculus , overlying which may be an encrustation of
modifications , cobble-ups, and jury -rigged inventions . The idea is
useful , therefore, to the extent that it reveals some simple computational pattern at the core of the immense complexity .
In the superior colliculus are a number of intervening layers, one of
which contains an auditory map that permits the animal to foveate
what it hears. In some animals one layer is a " whisker " map, which
enables the organism to foveate what its whiskers sense. As one
would expect, the neck muscles are also represented, so that in some
cases foveation is accomplished by moving both the head and the
eyeball. The principle of laminating suitably deformed topographical
maps and making vertical connections between them means that
phase space transformations .can be executed very simply and very
quickly . And the problem of neurogenesis looks more tractable the
simpler the computational principle .
Once this geometric technique of performing transformations is
understood , it will be observed that the maps representing phase
spaceson which transformations are to be made need not be literally
stacked one on top of the other . Maps can exploit the phase space
sandwich principle even though they are spatially distant from each
other , so long as the set of fibers connecting them preserves the
445
p
.y
.
Imm
Figure 10.15
(a) Projectiondrawing of Nissl-stainedcrosssectionof cat superiorcolliculusillustrating laminar organization. Dots corrEspondto cell bodiesof collicularneurons. (From
Kanasekiand Sprague1974.) (b) Retinotopicmap: a metricallydeformedtopographic
map of the visual hemifield, in rectangularcoordinates
, on the superficiallayer of the
right colliculusof the cat. Abbreviations: M, medial; L, lateral; P, posterior. (Adapted
from Schiller1984.)
relevant topographical relations . Additionally , it must be stressed
that even if we know that a cortical area is topographically organized ,
it is a further and more difficult question what phase space is there
represented . The more complex the nervous system , the more abstract will be the phase space representations deep in the neuroaxis .
Such abstract maps are what Konishi (forthcoming ) calls " centrally
synthesized maps ." Notice also that it would be simple enough to
add a,lamina representing real space (figures 10.16, 10.17) that could
be used in calculating the trajectory of a moving object .
The discovery that nervous systems abound in topographical maps
and relation -preserving interconnections (chapter 3) has long seemed
an organizational key that ought to help solve the functional mys teries . The phase space sandwich hypothesis is one possible way of
finding a fit between some anatomical facts and a computational
theory .3 It should be emphasized that the hypothesis suggests only
one way in which sensorimotor computations might be executed , and
other computational principles such as matrix multiplication , winner take -all strategies , and error -correcting strategies (see section 10.9)
may also have a place . Nervous systems likely use assorted tricks and
devices , and the phase space sandwich may be but one among many .
447
448
A Neurophilosophical Perspective
so
that
it
effects
the
correct
transformation
from
the
" motor intention vector," which issues from the higher motor centers , to the " motor execution vector ," which finally directs the mus -
matrix
matrix
in such
a fashion
that , after
identical
. In this
state
the
matrix
effects
a number
of iterations
the relevant
the correct
, the
eigenvectors
coordinate
are
transfor
mation . The authors further suggest that the climbing fibers of the
cerebellum- which had no role in their initial theory of cerebellar
function - constitute the special pathway by which such reverberative feedback modifies the transformational
lar network . Such a " relaxation " process constitutes a possible point
of contact with the " relaxation " algorithms advanced by the " connectionists
" in AI , whose
work
will
be discussed
in section
10 .9 .
, whether
or not
there
is feedback
. The
modification
and
novel sequencing of available skills to make new skills are the basic
ingredients of motor " learning ." This is not to say that new skills are
nothing other than constructs of old skill -atoms , since genuine
novelty does appear - skating is not just sliding one foot in front of
the other , for example. But it does suggest that the learning of new
skills draws on available skills and exploits error correction in order to
produce a novel motor program . Thus, for example, in learning to
pole-vault we begin by assembling a set of already existing skillsrunning , grasping, pushing the body , and so forth - and errorcorrect until a new smooth sequence is produced . It is in error
correction
that
feedback
becomes
crucial
Whether detailed explanations of learning within the tensor network theory will be forthcoming
remains
to be seen . It may
be that
the
tensor
network
449
others may be understandable only by invoking quite different princi ples, perhaps as an overlay on a geometrical analysis, perhaps in
terms of principles altogether independent of it . A vector-matrix
theory may, for example, give a good account of motor learning and
certain kinds of recognition
ing in humans .
in terms of modifications
to the geometry of
highly abstract phase spaces. Additionally , it seems to me quite possible that some capacities hitherto considered strictly cognitive may be
discovered to share fundamental elements with paradigmatic motor
and 9 .)
to do
it .
virtues of the
resort
, but
if we
are to know
whether
arrays
of neurons
fit the
tensor network hypothesis , ways must also be found for getting exacting detail on the response profile of a large battery of input
neurons, parallel fibers, Purkinje cells, and so forth .
At the moment , the neuron -by-neuron test is not available, because
450
A N europhilosophical Perspective
451
452 A Neurophilosophical
Perspective
The only theory of representations and representing we've got, runs
the lesson, is sentential . Critics of the sentential paradigm have often
been challenged to present an alternative theory of representing ,
where the implicit rider is that if the critic cannot say even what a
competing theory would look like , then he ought to keep mum . The
discussion of the tensor network theory so far has emphasized the
computationalpower of the theory in addressing the computational
questions of how the brain solves sensorimotor problems . But it is
important to see that it also shows great promise in the way it addresses the questions concerning how the brain represents
.
The tensor network theory opens a door, through which we can
begin to envisage a radically different paradigm in which representa
tions, even at fairly abstract levels of organization , are interpreted
as .points in .phasespaces
, and where tensorial transformations effect
.
transitions between phase spaces. Although some cognitive activity
probably is understandable as the manipulation of sentential
representations according to logical rules, many cognitive processes
likely are not . Indeed , the processes underlying sentential representation are surely themselves nonsentential in nature . What is needed
is a way to conceive of what nonsentential representing might be, and
the tensor network theory provides that much, even if , in the end, it
turns out not to be right .
Unbuttoning a bit , let us indulge in some unrestrained speculation
concerning how some cognitive representing might be understood
within the tensor network hypothesis . Consider the problem of recognizing faces, or the problem of how we can recognize a phonemic
string as the sort that could be an English word , even though it is not .
Both feats of recognition are accomplished routinely and effortlessly ,
yet both stubbornly resist a sentential analysis. Since a new approach
is called for , envisage instead the possibility that facial recognition
involves a feature phase space with as many dimensions as there are
recognitionally relevant features. The neurons keyed to respond to
individual features are the axes of the coordinate space, and the activity levels of the neurons jointly specify a point in phase space.
Suppose, to make it simple , that one neuron is sensitive to size of
eyes, another to shape of eyes, another to distance between eyes,
another to nose length , another to mouth width , and so on . There
may be several hundred dimensions to this facial phase space, and a
given face will therefore occupy a specific point in the phase space.
More likely I the system wants less precision, and two presentations
of the same face may have only approximately the same point in
phase space. The response patterns of the input neurons will determine where in the phase space the face is, and hence a face is repre-
453
sented when a given response pattern obtains . Notice that this view
precisely does not say there is a " grandmother neuron ," but says
rather that there is a response pattern of a whole set of neurons that ,
within limits , covaries with the presentation of Grandmother .
Aunt Bessie, and why there are perceived family resemblances. Un doubtedly certain features are more salient than others, and the geometry of the phase space will be warped to reflect this . Eyes, for
example, appear to be more important than freckles; thus, I can easily
recognize a photo of Einstein whether or not it has been doctored
with freckles , but change the eyes and the recognition problem is
more difficult . Presumably this is the sort of thing cartoonists exploit .
characteristic
too how small lesions can easily be compensated for , and hence why
loss of a small
isolated inability
(e.g ., Churchill
Suppose that
prototype face,
in their
454
A Neurophilosophical
Perspective
The photoreceptors responsible for color vision are the cones, and a
cone may have one of three types of visual pigment , with differing
sensitivity to specific light wavelengths . The " long-wave cones" are
maximally sensitive to light at wavelengths of 450 nanometers, the
" middle -wave cones" are sensitive over a large range but peak at
about
535
nanometers
, and
the
" short
- wave
area
of
sensitivity largely overlaps with that of the middle -range cones and
partly overlaps with that of the long-wave cones, have a sensitivity
peak at 565 nanometers .
dark
colors
are found
in one
corner
of the
lowest
level
, the
very
lightest colors are found in the opposite corner of the highest level,
and
so forth
455
The possibility that presents itself is that the phase space of perceived color has as its physical substrate a neuronal phase space,
whose axes are three streams of neurons carrying information about
reflectance efficiency values at the three wavelengths . The representa
tion problem for color perception , therefore, can be addressed thus :
colors are represented as points in a three-dimensional neuronal
phase space.
Moved by the hypothesis concerning color perception , one might
go on to speculate that the perception of tastes and odors and the
neural substrate of these perceptions can perhaps also be addressed
within the same general framework . Consider : the human gustatory
system appears to have just four distinct types of receptors for taste -
of distinct
have
axons, but they convey information across a synapse to the next stage
of neurons , whose response is substantially a function of the receptor
type to which it is specifically linked , though its response is also
affected to a subtle degree by activity in other receptor types . Now if
we think of each postreceptor neuron type as constituting one axis in
a four -dimensional phase space, then the activity of neurons on each
neurons
For example , suppose the salt receptor and the sweet receptor are
excited , but the bitter and sour receptors are not . The activity in the
for further resolution within the taste phase space. Accordingly , the
taste of Jonathan apples will be a position in phase space, different
from the position occupied by the taste of absinthe, and even in
minor degree from the position occupied by the subtly different flavor
of McIntosh
for odors , if , as it
seems, the human olfactory system relies for its perception of a vast
range of distinct odors on a finite number (perhaps just five ) of recep tor types .
456
A Neurophilosophical Perspective
theory (Rosch 1973, Rosch and Mervis 1975) is that many concepts
should
be
understood
not
on
the
model
of
lists
of
necessary
and
The possibility being explored is that we can understand the physi cal substrate fOl~ cognitive quality spaces in terms of phase spaces
defined by neurons at some suitable level of organization , and this in
turn may plough back insights concerning the nature of cognitive
quality spaces. At the very least, it is striking that a nonpropasitional
cognitive
with a geometrical
success in
modeling visual recognition (Hinton and Sejnowski 1983a), word recognition (McClelland and Rumelhart 1981), and memory (Willshaw
1981) using a vector -matrix approach . The computing basis for these
of units . Transformations are achieved by matrix multiplication followed by thresholding or some similar nonlinear operation .
Finally , it may be that the geometric approach w.ill not so much
compete with the propositional -logical conception of cognition as explain and absorb it . It is known that the resources of phase space
analysis are sufficiently powerful to model the structures and relations of logic and language. That is, logicians have shown (Van Fraassen 1970, 1972, Van Fraassen and Hooker
1976) that
linguistic
structure and the logical relations between them can be handled simply and elegantly within a geometric framework . Propositions , for
example, are modeled as subspacesof a logicalphase space, and infer ences
can
be
modeled
as vector
- to - vector
transformations
. Hooker
457
has suggested (in correspondence) that it is therefore possible to envision a unified cognitive neurobiology in which logic and the proposi tional attitudes , characterized in terms of phase spaces, vectors, and
matrices, fit in quite naturally with a wider geometric theory of representations and computation . In that event, the psychology of sentential attitudes and reasoning might reduce rather smoothly to
neurobiology .
Whether such a reduction can be achieved is obviously an empirical
question , but Hooker 's idea is that the general framework permits us
to begin to reconceive the structures and relations of logic and language and to exploit the geometric conception in developing models
of the relations between subcognitive domains and the propositional
attitudes . Thus, on this vision of things , the propositional paradigm
would be reinterpreted geometrically .
Even granting the power and elegance of the geometrical characterization of logic and logical structures, however , one should probably not be very sanguine about the smoothnessof any prospective
reduction of folk psychology . One important reason here derives
from the work of Stephen Stich (1983), which , as we saw in section
9.5, shows that the folk psychological categories of belief, desire, and
so forth , have internal deficiencies which entail that they must be
reconfigured if they are to be used by a scientific psychology . Stich's
view is that at best, reconfigured analoguesof these categories will be
used by a scientific psychology . If any sentence-like representations
are to be reduced in the manner envisioned by Hooker , therefore, it
will be these reconfigured analogues of folk psychological categories.
Some of this section is philosophical , in the sense that it flies some
theoretical balloons . I readily admit that a great deal of tough empir ical slogging needs to be done to take it from the realm of the possible
to the realm of the actual. But a major goal has been to respond to
those who assumed that nothing in a theory of sensorimotor control
can be relevant to philosophical questions concerning mental states
and cognitive processes. My summary answer is that as a result of
considering a theory of the basic principles of sensorimotor control ,
we can begin to see the shape of a new and powerful paradigm for
understanding computational processes executed by the mind -brain
and for understanding how the mind -brain represents at a variety of
organizational levels. The multidimensional phase space appears to
be a very powerful means for representation -in -general, and clearly it
is no longer acceptable to say that the sentential theory of representing is the only game in town . But this section is also philosophical
in the (perhaps different ) sense that it tries to limn the general shape
of a unified theory of the mind -brain .
458
A Neurophilosophical
Perspective
459
of content
rather
than
on
considerations
of location
may
" grandmother
be
selective
to
a range
of
stimuli
and
there
are
no
of connectivity
among sets of
than
electronic
devices
in the
execution
says ,
460
A Neurophilosophical
Perspective
If current
estimates
of the
number
of synapses
in
the
brain
are
critical
resource
that
is most
obvious
is time
. Neurons
whose
parallel architecture
distributed nature of information storage has suggested to some researchers that greater success in understanding
cognitive
functions
might be achieved by a radical departure from the sequential stereotype . The idea has been to try to understand how interconnected
neuron -like elements , simultaneously processing information , might
accomplish such tasks as pattern recognition and learning . As Ander son and Hinton have assessed the situation with respect to memory ,
If we abandon the idea that the basic method of retrieving
items
elements
to achieve
content
- addressable
mem
ory . (1981:10)
articles
. As
noted
earlier
, whereas
in a conventional
model
and
The
basic
notion
in all
these
models
is the
idea
of a state
vector ,
that is , that the currently active representations within the system are coded as patterns of activity simultaneously present on
the set of elements that comprise the system . (1981:17)
461
In a clear statementof connectionistprinciples- a sort of connectionist manifesto- Feldman and Ballard characterize the new
approach:
The fundamental premise of connectionism is that individual
neurons do not transmit large amounts of symbolic information. In stead, they compute by being appropriatelyconnectedto large num bers of similar units . This is in sharp contrast to the conventional
computer model of intelligence prevalent in computer science
and cognitive psychology . (1982:208)
Expressing essentially the same point , Von der Malsburg and
Willshaw say of connectionist models,
In computer terminology , the programmeresidesin the structure of
the local interactions; the result of the computation being the globally ordered state, which is arrived at by a parallel computation , and which is resistant to small alterations . (1981:83; my
emphasis)
The connectionists have abandoned the exclusively top-down approach characteristic of much research in AI and are looking for prin ciples to explain how elements in networks interact to produce global
results . The two types of approach need not be at loggerheads on all
counts, however , for it may be that serial simulations could be appro priate for some types of cognitive activity or for certain levels of
cognitive organization , even if not for all . Moreover , the problems,
strategies, and lessons, learned in the course of developing serial
programs to simulate cognitive processes, are there for the parallel
modelers to exploit should they prove useful . Let us now take a closer
look at what the parallel modelers are trying to do. The centerpiece
problem for organizing the discussion is the problem of how we see.
GlobalEffectsand Local Interactions
In order for an organism to see, its nervous system must be affected
by the world external to it . The fundamental fact constraining any
hypothesis about how a brain can have visual perception is that the
input to the visual system is the two -dimensional array of light falling
on the retina . Out of that stimulus array, the brain must concoct an
interpretation of what in the external world corresponds to the received pattern of light . And of course, there is no one inside to seethe
array and identify it as the sort of pattern made by, say, a bird or a
pineapple . There are just networks of neurons that interact with each
other and that , as a result of the interconnections , yield the global
effect that is an interpretation of the 2-D array . Since it cannot be
462
A Neurophilosophical Perspective
between input
and
There are many variations on the parallel model theme . For example, there are linear models (Kohonen , Oja, and Lehtio 1981) and
nonlinear models (Hinton 1981) of simple associative memory . Ad dressing mechanisms are replaced in some parallel models by hard-
463
to conventional
computers
in architecture
(parallel versus serial), in the method for data storage, and in the
nature of representation .
the
neural
substrate
of vision
is far better
understood
than
the
example, seems apt: 'I. . . , [it ] resembles memory in the way that a
toy glider resembles a bird . It does fly , in a rigid sort of way , but it
consists of two sets of binary units : detectors for receiving input from
the
external
world
, and
units
connected
to
detectors
and
to
other
false = 0), this can be understood as a sort of vote for the hypothesis
it represents. Final hypotheses are determined by fit with the image
and by fit with other hypotheses, so that the winning hypothesis
emerges after extended voting competitions
464
A Neurophilosophical Perspective
has a .8 probability of being in the I'true " state. In this sense, therefore , the
model
is stochastic
, a feature
that
enables
it to avoid
certain
problems (seebelow ). Note also that the relation between two units is
a matter of " synaptic" weighting and that such weighting imple ments constraints
synaptic weights can be changed, then the relations between hypoth eses thereby change, and the system is capable of learning . That is,
one hypothesis can acquire a higher probability given a particular
image by virtue of modifying the probabilities for the relevant unit 's
being in the I'true " state in those conditions .
on the basis
of local
interactions
, where
units
have
access
to the
responses of their neighbors and adjust their own responses accord ing to how their neighbors are responding . By iterative modification
is asynchronous ,
units
Once the detectors are activated by a stimulus , a range of connected units are activated , which means that a certain range of hy potheses are " in the game ," as it were . Some units activate certain
units
and inhibit
other
LEFT SUBNET
r
./' -.--
RIGHT
,
465
SUBNET
./' -.--
STIMULATION
INHIBITION
Figure 10.18
cube . The
network
consists
of two
interconnected
subnetworks
one
corre -
sponding to each of the two possible interpretations . (From Rumelhart et al. (1986). In
Parallel Distributed Processing
: Explorationsin theMicrostructure of Cognition. Vol . 2: Applications, edTJ. L. McClelland and D . E. Rumelhart . Cambridge, Mass.: MIT Press.)
467
tors connected negatively , and each unit receives one positive input
from the stimulus . To see how the system will run , suppose all units
are off , and then give one unit a positive value (input ). Eventually ,
the system will settle into a state where all units in one subnetwork
are activated and all units in another subnetwork are turned off . This
can be understood as the system's having relaxed into an interpreta tion of the line drawing either as a right -facing Necker cube or as a
left -facing Necker cube. Figure 10.19 shows three different states the
network might relax into , depending on what unit is first activated.
This cartoon network is not an example of how a visual system is
thought to solve the object orientation problem . Rather, it is included
to illustrate as simply as possible some fundamental principles of
networks for cooperative search.
If units in a network are interpreted as hypotheses, then the connections between the units can be interpreted as constraints among
the hypotheses . The input to the system can also be thought of as a
constraint ; hence, when the system is stimulated by an input , the
locally optimal state to which it relaxes is a state that satisfies as many
constraints as possible. Kirkpatrick , Gelatt, and Vecchi (1983) have
pointed out that there is a provocative connection between cooperative search and thermodynamics , inasmuch as the network and its
behavior also have a model in energy distribution . The central idea is
that we can understand how the network finds a good combination of
hypotheses on the model of a physical system settling into a minimal
energy state.
More specifically , Kirkpatrick and his colleagues have suggested
that we take as the appropriate metaphor the annealing of metals.
Given the proper cooling schedule, the heated metal will appropri ately anneal into the desired crystalline structure . In this case we
know that as the metal cools, highly improbable states are eventually
replaced by more probable states as energy is distributed and the
system comes to find (relaxes into ) the global energy minimum . The
idea is that in cooperative search, the system simulates annealing , in
the sense that parts find local energy minima and the system gradu ally finds the most probable answer- the global energy minimum .
Part of the value of the thermodynamic analogy, aside from its
intrinsic suggestiveness, is that it makes available .to the modeler the
range of laws, concepts, and equations in thermodynamics that he
can then use to explore and create new features and principles in
parallel networks . In other words , it imports a whole dimension of
systematicityfor free,9 as it were, and this systematicity provides an
exploratory framework . More than that makes the analogy appealing ,
468
A Neurophilosophical
Perspective
469
a
~
~
~
. ~
: Vh. ~
V/ h
~
~ ~
~
~
_ f
~
~
~ ~ -
~ , ~
- ,,~
I};/#
~
~:t/h
~ %
~
~ .J
~ YM
~ ~
~
~ r
~ J~
.%
~
~ l ~
' :Vh
. ~ ~ ~ r&
f-
- ~
~
~ ~
@ ~
@ ~
~-
%
~
~ ~
~
V/ h
'
~
~
~
m
///.
%~ %
~
~
~
~
- f~ri ~ -
- ~
~
~
~
~
~
~
~
~
? A : : Yk : : Yk : : Yk : : Yk ~
~
~
~
' //. ~
~
m ~
~ ~ ~ ~
~
~
mmmm
~
~
~
~
~
~
- ~
~
~
7h.
- -
~
~
~
~
f~
%f
~
~
~
- : %f
~ ~
@
L
- - = ~ - ~- ~ 1- ~
%
~
7h , ~
~
:m
mmPA
~ ~
~ ~ A
~ ~ .
~ ~ ,
% Pd
1
1
f~
~
@
f
~
~ -
- 0
~
v/~
~
~
~
-
~ %
~ ,J
;.-;~
,
j
~
~
~
~
~ rA~ ~ ~ ~
~ f
~m
~
~
~
~
7h ~
~
~
,J
~ ~
~ % m
V
//; ~
,~ .
~ ~ - '~
~
% ~% % -
@ ~///. ~
~ A~ ~ ,
@
7h
~
7h
7h, ~
% % % ~ ~
_ . %
~ ~
~ mmmm
: : Yk ~
: : Yk ~
: : Yk ~
: : Yk ~
: : Yk ~
: : Yk ~
~
~
: : Yk
~
d .
C
" .'
( ",
~ ~
, Z
,
, ~
,
, ~
~
' Z
Z
, Z
z
. Z
Z
, Z
" '"
v~
" '"
v~
'/ "
v~
Z
, z
"
' Z
" '"
v~
,." "/.
" ~ J?'-;:a: ~
~
", ~ %
z~
",/% "m
",-'z X1
"~" %
,,/.~ ~
","~ ~
""Z
~. ~, '~ ~
~ m
-'~ m
,~ ~
,~ %
, 'z ~'~ m
, ~' , ~ , ~ , ~ m
~ %
;;"~
~p~, ~
~~ ~
~ '~ %
~ :~ '~/ / , %
~ ~ ~
~:~;~ ~
~ft ~, %
~ ~ ~
~~~
~
~ ~
~~ ~
~~ ~
~~ ~
~ ~ z~ ;
r~~
/? ~
~~ ~
~
~
~ ~ %~ ~ ~ ~
.??} ~ m ~ ~
~~ ' l ~~
~' /
%
~h
;-;"
~;-;;, Ya
~ Ya
~h ~~
~
;-;"
WJ1
;-;,
~
~
;:;;~J
~
.
t% ~
%{
L;:;~z
~
;~- ; ?
"W
, , ; -l
'1
~%~
' ;'-;/
~
~
~
~ :/ J~
~ % ~ ~ ~(
~,,'~ z~' ~ ~
~;
~:
~ ~ ~
r~~ ~
~
%
~
~
~
~
~
~
~
~
~X
~
~;
~
Z ;~{ ~
~ ; ,l
~
~
~
~~/
~
~ ~
~ ~;-: ~
~ x"
r\/,,
' / ;.,
;;'";'A '~/~f
~
~ ' ~A ~
~
~
-'
~~~ ~
r7 ~,A
~
~ ~ ~
%
~ f ::-~~
~ ~ % ~ ~ ' l , ~ z ~ ;-1
~; ~
,,~
?~
~
~
~
r. / "
, ,/ ~
'
'
/ ,, '
~ ~ z~ ~ ~
~
w
~
~
@
~
~
~ ~
~~
'/ '
, ,' "
V$ / A '~X
,,
9;~ '~/ !~
~~ ~
~~
~
~
~
~
~
:!~z ~
~ -;., ~
~z
~
::-'~ ~
~z
%,;,
~' A ~ ~ ~
~
~
~ ~ ' /~ ~ ~ ,/, ~ ~ ~
%
~r~'.~,/ ,' .' '"
Z Z
~
~
~
~~ ~ ~ ~
~
~
m
"'/,.~
~
~
~
~
~
~
~
~ ~
Z 'Z
"" ,
/"/.
/ '
~ ,
~ ", "Z~ Z gz
~
m
M
" ,//,~~"
, -"', ,z' Z ~;.-' %,,~' / /, ~Z " "v,~
~z I
" '"
r"/" "
~"
%~ ~
' " ~ ' ~" ,,/ 'z ~
~/ "
~
Z, ' " m
/,"~ m
"'"" z ~.
~,~ //"
~~
"" Z
~ ,." ~
~
~
~
~
~
@ ~
@ ~
%~ ~
%~ ~
;;::;; @
~ ~ ~
%~ @
~ ~, ~
~~
, '.
~' /~
/ ':{
M' ~
' Z ~" .," Z" '~' ///, ~' %
,~ .~
~" ~ ~
., ~ z ~Z ~
r~
,/ " '' : -;
"
~
'Z
%
' ~
i' '' ""'~ " "~' , %~' ,~
,
~~ ~
~
~
~ z
"%
,, '
// /" "
~ ,,~, ~
"""~
,z ~
,,~ m
,z ~
V/ ~;-: ~
: : Yk
~ ~ ;;::;;
~/ A ~
~~;-: ~
~~ ~
~~
r'//
z ~
~~~ ~( /}A %
~~~ ~~/A
~
} ~/A
~ ~ ::Yk
~ ~} WA
%~
~
~~
~ :., zz
~ ~ ~ ~ ~
~h
~
~%
~ ): ~
(~~ ~
~ ,,: ~
~ ", m
~ " '~7
~
/ ~
/ /-J
1
~~ / ~~':';
~p ;.-;"
~ ~~~/ , ;:/-J
'/ m
~/,
~~ ~
~~~ ~
~ ~~ ~
~/
~
rp
~
~~ ~
~~ ~~ ~ z ~
.-' ,
, ,/, - : ~
" / r,/
", '
.~ '~
~~, ~, ~'
,,
%
,,~ %
. " , ,z.. '~, 7,
, : ""
Y~ Z
,.
~, Z ~~'" m
'", z ~" ,~ "-"
~. /. ~'~ %
".,,//.
; -/
';;
"
/ . /~
// '
" / ,,
r/: '
,,-', ~
z;.-" ~
"",,~ ~
/,"~ "%
"'~ "~. Z //.
~~ ~, / ~
uZ ~
","~ ,~, ~ 7.-~
~ ~
M
/",,~ m
"'/, Z Z
%
o"'. ' . ~
z'~ z',
,~ ~
,/.
~ m
:'~ m
~Z I
' '
~
/
/' "
r/ "
~~ ~'Y; ~
'/ ~ y ;
~
,,
'
, ,"
'/ "
y/
~ ,~ r ,' / , ~ Z ~ :,-: ., y ;
~~
~x
~~
?;~
r: ~
~ ~
~ ~
~ ~
I ' :' l
~ ~
(~ /
Z /
::::C
:-1 ~
~ z
~ ~
~ ,'
, ,' "
~ ,'
~" 1 ~
7
/.
~;
/ / ~
" ', z
/ ' ,
"
"
::::
~~
7ft
%~
7- /
7 , '
~ ;,
% /
~ ~
~ /
~ /
% ~
7ft
; ,~/
7 ,'
~ -1 %
;;:x
' /
h /
~ ~
~~
~~
,
,~,"rz
~ 'Zz
r '"'
,/ '/ , ~
/z/, ..,
'
// /,/:"'
/"/ /"
'i'/
/
/ .'"" %
~
Uz
~
~
~
f1
""/,. '/
/"
/"" , '~'., /,,, -"/'"' ~
"
r/ '
''
(/ 1 , 7~ ~~
/ -,, '
~/ '
'
~ Z
Z , ~"
~ ~
~ z
' //.
7 . -' l
~ ~
' // / .
1-
/ "Z
~ ~/
~; ;:: ~
~
~ -;., I ' : ~
~p
~ /.
%~./" , mz
7/V/
" '// ~'
,",'/
-y,,~'
/"/v /~
"z
'.. /,:-:
-:
z '" mz
"" . mz
~
m
~
7Z "/ :
" ,,/
' -", ~
"? ~
' ~'
Z . /"
Z
r"":'/ " ~'
" ~ i'
Z/
-;. ~
/z,,,
Z-' ;, :-:U{ 1 m.
:-: Z~
/ , ~"/ %~
/ / .A
Z ~/ %
~
Z"",; "%;./ -'X~ "~
/ , '~" //:",%
~
;,
/ / /.
~~ %
-'
~ ~ ~~~;~ ~~
~ ;., @
~ /.
r ;' %dz %
~'-1
/, ~
~~ ~
%/. ~
~
'~/// , ~
""'/,,~
' ,%
r"'"
~, ;.- ~
Z
m
/" m
;.-, %
U',..' :, ~
"~"
Figure 10.20
This diagram shows three successive stages in simulated annealing as the network
solves the problem of separating figure from ground . (Hatchmark squares ~ are figure
units , .... are edge units .) Physical constraints , such as that edges tend to be continuous
and that figures are internally connected, are built into the algorithm . (a) shows the
network at " high temperature ," where the figure and edge units make a structureless
pattern . In (b ) the network
ture appears as figure units support edge units whose orientation is consistent with
them . (c) At " low temperature " local inconsistencies
" crystallizes" to the correct solution , and the block letter " c " appears as figure against
a bare background . In (d) the center of attention shifts to the outside and causes a
figure -ground reversal. (Courtesy T. Sejnowski . Based on Sejnowski and Hinton
(forthcoming ).)
470
A Neurophilosophical Perspective
superficial minima , so that , after a while , the lower a minimum is, the
more likely it is that the system will be there (Ackley , Hinton , and
Sejnowski 1985) .
of an
algorithm concerning average energy distributions devised by Metropolis et al . (1953) . The effect of the addition is that noise is added to
the decision rule . Occasionally, therefore, a unit will adopt a hypothesis that is less consistent with the hypothesis gaining strength
in its neighborhood , and the locally improbable hypothesis could
eventually win global status depending on how it comports with
hypotheses in other neighborhoods . On this view of matters, noise in
the system turns out not to be an incidental and ignorable property of
the system , but to be important
function . As Hinton
noise in networks would be exactly the sort of thing one would miss.
And yet it may be exactly the sort of thing that needs to be considered
in order to get an adequate explanation of cognitive processes.
Do Real Neuronal Systems Work This Way ?
tion , recall that because the machine that implements the program
has a parallel architecture , and because the units are neuron -like ,
471
with connectivity grossly patterned to simulate neuronal connections , the Hinton -Sejnowski model is to a surprising degree biologi cally constrained . Therefore, it is not merely away , an any-old-way ,
to do pattern recognition . More specifically, it is a way for a set of
simple units , organized in parallel , with rich cross-talk potential and
interaction constraints such as " synaptic weights ," to do pattern
recogni tion .
For the model to provide a theory of certain brain functions such as
pattern recognition , it will need to account both for fine -grained anatomical and physiological data and for higher -level psychological
data. As in the case of the tensor network theory , we shall need to
evaluate how well it explains and unifies the data, how it can be
tested, and whether it is rich enough to suggest new experimental
tests. To be fair , we must bear in mind that the parallel modeling
strategies are of very recent vintage and that what we see thus far is
really just a beginning . What we have is not so much a theory of how
the visual system functions , or even how one aspect of it functions ,
but rather an approach that may lead to a theory of the phenomena .
There are anatomical and physiological facts about the visual system that at some suitable stage need to be embraced within the
model , assuming it is adequate enough to be worth extending - for
example, the center-surround organization of the retina, the X, Y,
and W ganglion -cell pathways with their distinct properties , the retinotopic maps in distinct areas of the cortex, and the characteristic
columnar architecture and connectivity patterns . Some researchers
have taken steps in this direction . For example, Zucker and Hummel
(1983) claim that their relaxation model is faithful to and accounts for
both the processing in the X-pathway and the center-surround organization of retinal cells.
Certain psychological features of vision seem to be difficult to account for within computer vision models. For example, subjective
contours are a puzzle because there is nothing on the 2-D image array
that corresponds to the subjective edge and contrast in color perceived. It may be that these features are not the outcome of low -level
vision modules , as David Marr has claimed, but essentially involve
the contribution of higher centers, as for example, Irvin Rock (1983)
claims. Marr (1982) hoped that he could elicit subjective
contours in
his computer model as part of the normal process of edge filling -in ,
but Rock points out that many casesof subjective contours cannot be
handled that way and need a figure -ground distinction . Such a distinction may need higher -level /lhypotheses" to solve the problem for
the lower -level module . The recent finding by von der Heydt ,
Peterhans, and Baumgartner (1984) showing that there are cells in
472
A Neurophilosophical Perspective
area 18 that can be driven by subjective contours indicates, nonetheless, that ostensibly high -level business can be investigated neurobiologically , and this may encourage the connectionists . Many other
features also suggest " intelligence in perception " to Rock, such as
size constancy and orientation constancy.
Although Marr 's particular computer vision model does seem vul nerable to these objections, the Hinton -Sejnowski model may have
the flexibility to accommodate the intervention of high -level hypoth esesin visual perception . One can envisage stages of relaxation in the
visual system, such that a hypothesis submitted by one level could be
rejected at a higher level, with instructions sent below to rework the
data and, subject to certain constraints , to come up with a better
hypothesis . In other words , the relaxation process could sweep up a
wider and wider range of " hypotheses," and the stochastic nature of
the processeswill permit initially improbable combinations of hypoth eses to win out . This conception is compatible , of course, with the
view that vision is often subject to descending control .
In a recent paper, however , Sejnowski and Hinton (forthcoming )
propose a cartoon story to indicate how a network can solve figure ground problems by relaxing to a best-fit hypothesis . The model was
designed so that missing line segments could be filled in , and a shift
in an attentional / spotlight " permits the network to relax into a new
hypothesis specifying what is figure and what is ground . (On the
attentional spotlight , see Crick 1984and section 10.10.) The successof
this model in handling the faces/vase caseof figure -ground ambiguity
indicates that at least some figure -ground solutions are well within
the capacity of cooperative search.
A further question concerning the neurobiological fidelity of the
model arises in connection with the basic units . Recall that these are
simple binary units , whereas real neurons are not simple on-off units .
A neuron has a spontaneous firing rate that can be increased by
depolarization and decreased by hyperpolarization . Its firing frequency can be affected by a large assortment of factors apart from the
current changes caused by input neurons (chapter 2). Extracellular
circumstances, the recent history of the neuron 's behavior , and the
intrinsic properties of particular neurons will bear upon its spiking
profile . The graded potentials induced at postsynaptic regions also
have a complex profile not captured by the simplifications of the
computer model (Freeman 1975). The question , therefore, is whether
the simple units of the model can be translated into the more complex
neuronal units that actually exist.
The optimistic answer is that networks of binary units engaged in
cooperative search are rather like the cartoon story of Roger, in the
Theories
of Brain Function
473
sensethattheysimplify,idealize,andapproximate.
Butif the basic
principles
oforganization
areright,thentheycanbefancied
up to
makethecartoon
storymoreclosely
resemble
therealstory.Addi-
tionally,
Hintonhaspointedout(personal
communication)
thatifone
regards
theoutputofa neuron
overa veryshorttimeperiod,
suchas
a fewmilliseconds,
it reallyis a stochastic
binarydevice.Thatthe
firing
rateoverthelonger
timeperiod
isfundamental
seems
unlikely
Additionally,
it is notclearthatthemodelis fastenough.Relaxationtechniques
do havegreatappeal,butcross-talk,
voting,and
revoting
taketime.Torepeatan earlierpoint:if a recognition
re-
sponseoccurs500msecafterthestimulus,thereistimeforonlyabout
twentysynapticrelays;if it hasa 1,000mseclatency,thereis timefor
onlyaboutfortysynaptic
relays.
Moreover,
theresponse
latency
includes
thetimetakentomobilize
andexecute
themotor
response,
not
justthetimetakentoreacha consensual
decision
oncategorization
of
theinput.Still,as Hintonhasremarked
(personal
communication),
thereis timeforabouttwentyiterations
(individual
spikes)
in 100
msec,soperhaps
timeconstraints
arenota problem.
I donotsaythat
thesemodels
aretooslow,butonlythatspeedisa majorconcern.
Finally,
wemaywanttoposetoconnectionist
models
ofpattern
recognition,
memory,
orlearning
thefollowing
question:
Howarethe
proposed
solutions
meantto fitintothewiderstoryof thebrains
Putcrudely,
brainsarenotinthebusiness
ofpattern
recognition
for
itsownsake,andthenatureofpattern
recognition,
asaccomplished
achieve
motor
control.
Evolution
being
whatitis,pattern
recognition
istheretosubserve
motorcoordination.
Andwhatgoesforpattern
recognition
in thisregardgoesalsoforlearning
andmemory.
Since
connectionists
withvisionmodels,Whathappens
next?Howare the
resultsofa relaxation
algorithm
to figurein thetheoryofhowthe
organismmoves?Aretheresultssuchthattheycanbeusedin motor
control?
Byraisingthesequestions
I donotmeanto implyeitherthatcon-
nectionists
areunawareofthemor thattheywillbeunableto answer
them.Rather,theimplication
isthatifweignoremotorcontrolasthe
context
withinwhichwetryto understand
patternrecognition,
we
474
A Neurophilosophical
Perspective
puter , of course, does not have to feed itself , build a nest, mate, or
hide; hence, it can afford idly to deploy pattern recognition , platon ically unconcerned with matters motoric . The brain , making its living
in a rough world , red in tooth and claw, cannot afford such purely
contemplative luxury . The point , therefore , is that insofar as evolution solved the problems of sensory processing and motor control
simultaneously , we may find it profitable - nay, essential- in shaping
our theories, to mimic evolution and aim for simultaneous solutions
as well . (See also Llinas (forthcoming ) and Foss 1986.)
10.10 The Neurobiologyof an Attentional Operation
The final example of a theoretical development to be discussed here
has as its major target the neurobiological mechanisms that subserve
visual attention , and its author is Francis Crick (1984).
The background problem for this theoretical venture must be
understood first . A fundamental puzzle concerning visual perception
arises from the fact that cells in different areas of the cortex appear to
be specialized to respond to distinct dimensions of the physical
stimulus . Cells in some areas are maximally responsive to movement ,
others to lines in specific orientations , others to conspecific faces,
others to colors, others to stereoptic disparity , and so forth (chapter
3). In oth ~r words , perceptual features seem arranged by topic and
dispersed hither and thither around the visual cortex. Our perceptions , by contrast, show no such disunity . One sees a unified composite, such as a running black dog or a falling yellow ball . We seem
driven to the inference that there must be neuronal means for conjoining distinct properties - yellow color and roundness- such that
w hat is seen is a yellow ball .
There is an additional important element in this puzzle . Although
we appear to be able to perceive an indefinite number of distinct
combinations , there cannot be hard-wired in distinct cells corresponding to everyone of these combinations because there is
insufficient neuronal hardware for that . Therefore, Crick suggests, in
order to express the particular conjunction of properties in the
stimulus , there must be temporary associations of cells, where each
represents a property from a distinct dimension . The association is to
be temporary so that a set of cells can on one occasion partici pate in
the " yellow ball" combination , on another occasion a " yellow box"
combination , and so forth .
To a first approximation , then , we can expect the visual system to
generate perceptual combinations by temporarily associating in some
manner sets of appropriate cells, where the association has down -
475
need to
be so formed . Some, of especial biological significance, may be hardwired in genetically , and others, subserving the perception of words ,
for example, may be permanent cell assemblies resulting from " overlearning ." Even granting large numbers of such permanently associated cell groups , it is evident that many associations expressing
red H and a green T. Treisman has also found that a subject searching
a presented card for a single feature- say just the H-shape (color
does not count )- takes roughly about the same amount of time regardless of the number of items, whereas when he is searching for a
color-shape conjunction - say, a green H - his search time is a linear
function of the number of items in the array . Speaking loosely, the
feature seems to " pop out ," whereas finding the conjunction requires
attentive
1980. )
scanning
tive " searchlight ," moving from one item to the next, until the sought
item has been found . The searchlight is not the only attentional mechanism exploited by the nervous system. Also in evidence, for example, is an attentional effect when one's eyes saccade to a
movement in peripheral vision or when one' s head moves to foveate
the area from which a loud sound emanates or where one expects to
476
A Neurophilosophical
Perspective
.-.
'
~.-...~_...~
- - -- ---~ e~ '
"
titular
mplex
..I
::::&;::::::::~: am
us
Figure 10.21
The main connections of the reticular complex, highly diagrammatic and not drawn to
scale. Solid lines represent excitatory axons. Dashed lines show inhibitory (GABAergic )
axons. Arrows represent synapses. (From Crick 1984.)
477
again for another 80- 150 msec. One possibility , therefore, is that
neurons of the reticular complex respond to incoming activity by
giving the thalamic neurons a hyperpolarizing
have
ceased
, and
so on in
sequence
the
. He _calls
two
cell
sets
are
connected
synapes
that
alter
in
the
is the production
of rapid
cortical regions .
conjunction .
light to do its work and hence for the relevant cell assemblies to form .
Crick is quick to point out that the theory is of course imprecise in
478
A Neurophilosophical Perspective
479
SelectedReadings
Ballard, D . H . (1986). Cortical connections and parallel processing: Structure and function . Behavioral
and Brain
Sciences 9 / 1 .
Llinas, Rodolfo R. (forthcoming ). " Mindness " as a functional state of the brain . In Mind
matters , ed . C . Blakemore
and
S . Greenfield
E . Rumelhart
. Oxford
: Basil
Blackwell
Mass . : MIT
Press .
of functional
geometries
system . Neuroscience
16 :245 - 273 .
Rumelhart , David
Press
Chapter 11
Closing Rell1arks
At this stage in their history the brain and behavioral sciences are
monumentally exciting , for we appear to have embarked upon a period when an encompassing scientific understanding of the mind brain will , in some nontrivial measure, be ours . Theories- of thp
large-scale, governing -paradigm , unifying -framework kind - are beginning to emerge, and they will evolve and come to structure both
the research enterprise and, undoubtedly , our very way of thinking
about ourselves. And it would be amazing if the new theories and the
new discoveries did not contain surprises of such magnitude as to
constitute a revolution in understanding . In its power to overturn the
" eternal verities " of folk knowledge , this revolution will be at least
the equal of the Copernican and Darwinian revolutions . It is already
evident that some deeply central folk psychological concepts, such as
memory , learning , and consciousness, are either fragmenting or will
be replaced by more adequate categories.
Those who suppose that science and humanism must be at loggerheads will greet this forecast of the future with no enthusiasm .
They may tend to see the revision of folk theory and the rise of
neurobiological -psychological theory as the irreparable loss of our
humanity . But one can see it another way . It may be a loss, not of
something necessary for our humanity , but of something merely
familiar and well -worn . It may be the loss of something that, though
second nature , blinkers our understanding and tethers our insight .
The gain,. accordingly , may be a profound increase in the under standing of ourselves, which , in the deepest sense, will contribute to,
not diminish , our humanity . The loss, moreover , may include certain
482
A Neurophilosophical Perspective
folk presumptions and myths that , from the point of view of fairness
and decency, we come to see as inhumane . And among the desirable
losses may also be numbered certain widespread and horrible diseases of the mind -brain .
Divesting ourselves of such things as these is clearly in the spirit of
both humanism and science. In any case, it is a mistake to see science
as standing in opposition to humanism ; rather , it is the political and
entrepreneurial abusesof scientific knowledge that have displayed
a catastrophic disregard for humanistic principles . And we shall
certainly have to guard against similar abuses of neuroscientific
know ledge.
A primary objective of this book has been to show that neuroscience matters to philosophy . I have advanced three general lines of
argument in support of this view : (1) mental processesare brain processes, (2) the theoretical framework resulting from a co-evolution of
neuroscience and psychology is bound to be superior to folk psychology, and (3) it is most unlikely that we can devise an adequate theory
of the mind -brain without knowing in great detail about the structure
and organization of nervous systems.
The correlative theme is that philosophy matters to neuroscience.
My strategy here has been to sound that theme less by explicit argument than by implicit demonstration . The central point , nevertheless,
is that neuroscience needs philosophy because ongoing research
must have a synoptic vision within which the immediate research
goals make sense. Such a synoptic vision , transcending disciplinary
boundaries but informed by the relevant disciplines , testing the integ rity of the governing paradigm and investigating alternatives , is phi losophy . At least, it is one very traditional way of doing philosophy .
But this sort of philosophy is not an a priori discipline pontificating
grandly to the rest of science; it is in the swim with the rest of science
and hence stands to be corrected as empirical discovery proceeds.
Moreover , discoveries in neuroscience will undoubtedly change
out of all recognition a host of orthodoxies beloved in philosophy .
Barring a miracle (or a calcified stubbornness), it will in particular
transfigure epistemology , as we discover what it really means for
brains to learn, to theorize , to know , and to represent. Neuroscience
may even teach us a substantial thing or two about how science and
mathematics are themselves possible for our species.
So it is that the brain investigates the brain , theorizing about what
brains do when they theorize , finding out what brains do when they
find out , and being changed forever by the knowledge .
Notes
Chapter 1
1. The word " neuron" is Greekin origin, and the Greeksused it to mean " nerve"
(large, visiblecable), though they sometimesalsousedit to mean" tendon," which
makesfor confusedreading.
2. JamesBogenhas reminded me of the delicateand difficult aspectsof Golgi staining, appreciationof which makesthe reticularist's positionseemmoreunderstandable than it otherwisemight.
Chapter 2
1. " Process
" seemsat first a rather odd word for structures, applying asit commonly
doesto a train of events. But it is the standardword and I shalluseit hereafterin its
neurobiologicalsense.
2. For instance, if you carefullyremovethe soft white pieceof spinalcord in the cleft
of the vertebralbone of a lamb chop and cut through it cleanly, you will seethe
centralgray (approximatelyin a butterfly shape) surroundedby white matter.
3. Receptorsfor the benzodiazepines
(e.g., Valium and Librium) havebeenfound in
the brain (discussedin Braestrupand Nielsen 1980), as well as a peptide that
naturally occursin the brain, actsat the benzodiazepinereceptor, and increases
anxiety (Ferraro, Guidotti, Conti-Tronconi, and Costa1984
).
4. This is true of mammals, wheremaleshavean X and a Y chromosomeand females
have two X chromosomes
. However, in birds the femalesare heterogametic(XY)
and the malesare homogametic(XX), and in somereptilesthe sexof offspring is
determinednot geneticallybut by environmentalfactorssuchastemperatureof the
incubatingsand. If you area snappingturtle, you will be femaleif your motherlaid
the eggyou are in on sandwhosetemperatureis at the extremelow of 20 degrees
Celsiusor the extremehigh of about 30 degreesCelsius; you will be male if the
temperatureis somewherein between(Harvey and Slatkin 1982
).
5. The hyperstriatum ventrale, pars caudale(HVc), and the robust nucleusof the
archistriatum(RA).
Chapter 3
1. Eachneuron has a favoredsound frequency- calledits " characteristic
" or "best"
frequency- at which it maximally respondseven when the sound is very faint.
Typicallyit will respondto a rangeof slightly higher and slightly lower tonesif the
decibelsare increased.
484
Chapter 4
1. Neuropsychologists(Squireand Cohen1984) havesuggestedthat the phenomena
invite a distinction betweendeclarative
and procedural
memory, where what is declarativeis in somesenseaccessible
to consciousawarenessand what is procedural
is not. There are some animal models using monkeys(Zola-Morgan 1984) that
confirm the dissociabilityof the two capacities
, but since the monkeys do not
literally " report!' anything, this basisfor the distinctionis acknowledgedto be still
a provisional labeling of the assortmentof things amnesicsrespectivelycan and
cannotremember. To this extent, it is essentiallya behaviorallabeling, rather than
a theoreticallyembeddedcategorization
. (Seealso section9.4.)
2. A lesionis any damageto nervous tissue that impairs normal functioning of the
neurons. For example, a lesion could be causedby surgery, disease
, a tumor,
vascularabnormalities
, drugs, pressure, or a blow to the head.
3. The problem is with what Kolb and Whishaw (1980) call " the n = 1 studies."
4. Spencerarguedthus:
If there is someorganization, it must consistin that same" physiologicaldivision of labour" in which all organizationconsists; and there is no division of
labour, physiologicalor other, of which we haveany example, or canform any
conception, but what involvesthe concentrationof specialkinds of activity in
specialplaces(1855:611).
5. Grunbaumchangedhis nameto Leyton in 1915.
Chapter 5
1. What seems phenomenologically unappealing about the Crick-Mitchison theory is
that it doesn't appear to account for regularities in dream content . For example, in
adolescence (at least) many dreams are of a sexual nature , and from adolescence
onward people also report countless variations on the " test" dream. Freud's
theory , unappealing on other grounds , does have some strengths on this issue. In
support of the Crick-Mitchison hypothesis , however , is the finding by Hopfield ,
Feinstein, and Palmer (1983) that connectionist models quickly suffer from overload unless there is a " forgetting " proced ure .
2. There is some controversy concerning one specific tracer technique developed by
Sokoloff and his colleagues. The Sokoloff technique uses 2-deoxy-glucose, an analogue of glucose, and the questions pertain to the basic biochemical and metabolic
assumptions of the technique . (See Fox 1984 and a response by Raichle and Ter-
485
in taxicab
number
1729 and
remarked
that
the number
seemed
to me
rather a dull one, and that I hoped it was not an unfavorable omen . " No ," he
replied , " it is a very' interesting number ; it is the smallest number expressible
as the sum
of two
cubes
in two
different
ways ."
"'
5. The ventricles are cavities in the brain filled with cerebrospinal fluid (figure 3.6).
Sometimes the canals connecting the cavities are blocked and the cerebrospinal
fluid cannot drain out but continues
(See section 3 .2 .)
Chapter 6
1. I realize that this caveat may not do me much good . In trying to keep the discussion
simple, I will inevitably slight many a nicety and outrage some historian or other .
Philosophers
are notoriously
sensitive , perhaps
of
their subject, and for some, every nuance and every shade must be honored . My
choice was simple : either run the risk of offending those historians who find any
simplification a brutal oversimplification , or omit altogether the history relevant to
the wider project . The choice was obvious . As usua~, selected readings will be
provided at the chapter end .
2.
Mill was the exception to this empiricist principle , since he thought even mathematical principles were , when you got to the heart of the matter , empirical .
3 . There are , or course , many versions of idealism . The common
theme is the view
that the fundamental reality is mind , not matter . Furthermore , idealism typically is
antinaturalist
in the sense that it is opposed to the view that the operations of the
mind can be understood in terms of causes and effects . Nineteenth -century ideal -
ists also adopted the view , owed to Kant , that experience is not merely " given" but
involves some sort of organization
, categorization
, and interpretation
. This concep -
tion of the mind as actively contributing to experiences rather than passively recording it figured in the idealists' opposition to naive realism, which says that our
perceptions closely resemble the external objects perceived. Hegel was the domi nant figure in the idealism of the later part of the nineteenth century . A noted
British idealist of the 1920swas J. M . E. McTaggart, who wrote in rather mystical
fashion of the unreality of space, time, and matter . It was this thesis that
scientifically attuned philosophers found particularly offensive and worthy of
ridicule
4.
Though ardor for this sort of argument has visibly cooled, it is by no means dead.
For a recent example, see Anthony Kenny , Master of Balliol College, Oxford , scold
Richard Gregory for having fallen into fallacious ways of talking about the visual
system (Kenny 1984).
Chapter8
1. I explore these and other examples in considerable detail in the aforementioned
paper (1983).
2. One argument for substance dualism (see Eccles 1977) derives from certain
neurophysiological results obtained by Benjamin Libet et al. (1979). The argument
is particularly interesting because it is perhaps the only argument based solely on
objective empirical data. According to Ecclesand Libet, the data show that a mental
event precedesin time the brain states causally responsible for it . In their judgment
this can best be explained in terms of the nonphysical mind antedating the experience in time . As Eccles and Libet se,e it, if the mental event and the physical
coun terpart are not cotemporaneous, they cannot be identical .
After analyzing Libet's papers, I came to the conclusion that the interpretation of
the data was unwarranted and that other simpler and more straightforward inter pretations were readily available. In brief , I could not find any grounds for conclnding that mental states are distinct from brain states. The experimental setup and the
interpretation of the data are too complex to present here, and for an account of the
disagreement between Libet and me on this question I refer the reader to Patricia S.
Churchland 1981a, 1981b, and Libet 1981.
3. Others who have explored this view , either critically or sympathetically , include
Robinson (1982), Campbell (1983), Margolis (1978), Dennett (1978b, forthcoming )..
Shoemaker (1981).
4. Here is a simple way of summing up the difficulties by drawing on distinctions
used in philosophical logic . Either the contexts of the premises are transparent or
they are opaque. If they are transparent 1 then the second premise is justified only
by begging the question . If they are opaque, then the properties are not extensional, Leibniz 's law does not apply , and the arguments are fallacious. Either way 1
the arguments fail . I suspect that part of the appeal of arguments (A ) and (B) lies in
their being readable either as transparent or as opaque. Just as one pins down the
error on one reading , the other reading, like a Necker cube, flips into focus and
seems to be the intended reading . And so it proceeds, flipping back and forth
between the criticisms .
5. For a similar version , see also H . Robinson 1982.
6. For a thought -experiment that indicates how it is conceivable that our heroine
487
might use her utopian neuroscience to introspect the state of her mind -brain , see
Paul M . Churchland 1985.
7. For a different view on these questions, see Stich and Nisbett 1980.
8. I owe this argument to Paul M . Churchland .
9. This argument bears a superficial similarity to Searle' s (1980) argument , but it is
importantly different . As I understand it , Searle's argument contains a crucial
qualification in premise (2), namely that no computer model that mimics mind brain capacities only at their highest level of organization can be adequate, though
one that mimicked " all the causal powers of the brain " would be. Searle is a
physicalist , and his argument is directed not against reductionism but against what
he calls the " formal models" in artificial intelligence and cognitive psychology . It is
not altogether clear what is embraced and what is excluded by " formal ," but if
" purely formal " is roughly equivalent to " sentential," then there is some agreement between us since I am critical of the sentential paradigm . (Seechapter 9.) If it
is not equivalent or roughly so, then perhaps some of my objections will ruffle
Searle' s argument as well .
10. For a more fully developed account, see Paul M . Churchland 1979, Stich 1983. For a
different account, see Dretske 1981.
11. Nor is it clear that we need to care very much if what we want is a science of the
mind -brain , since the assignment of content to internal states may not figure in the
generalizations that describe information processing in cognitive organisms. For
arguments in support of this view , see Stich 1983.
Chapter9
1. So far as I can determine , Sperry's (1976, 1980) conception of mental states aligns
itself more with functionalism than with anything else, though his arguments
against reductionism seem confused to me. To the extent that I can understand
them , they seem to have a family resemblance to those considered in this and the
subsequent section.
2. Or perhaps some subdomain of cognitive psychology . Exactly what categories and
generalizations are at issue here is not very clear, but the intent , at least of Pylyshyn and Fodor, is to include the whole of cognitive psychology .
3. Perhaps these are the sorts of considerations that lay behind Searle's (1980) criti cism of functionalism .
4. This is just a partial specification of what is needed; for Hooker 's complete account,
see Hooker 1981, part III . The rest is not unimportant , but I think it need not be set
out here.
5. For a discussion of the opposite view , see Patricia Kitcher 1980, 1982.
6. Something like this argument may also be what moves Sperry (1976, 1980).
7. There are other variations on this theme to the effect that meanings are irreducible
to physical states, but I take the presented version to be the most coherent and
most persuasive. Davidson (1970) most particularly has proposed antireductionist
arguments , and though his work has been widely discussed in the philosophical
literature , I find Dennett 1978b, Fodor 1975, 1981, and Pylyshyn 1984 clearer and
deeper. For a Davidsonian attack on the position I defend , see Horgan and Wood ward 1985. For yet other arguments defending antireductionism , see Margolis
1978, C. Taylor 1971, Vendler 1984.
8. This idea has been intensively explored by Dreyfus (1979, 1981).
9. As I understand it , the frame problem refers to a rather more circumscribed prob -
Chapter10
1. Pellionisz and Llinas are by no means alone nor the first in taking representations
to be vectors in coordinate spacesand computations to be vector-to-vector transfor mations . For an early application , see Pitts and McCulloch 1947. The vector-matrix
approach was also used by Rosenblatt 1962 in the perceptron work . For recent
examples, see Kohonen et al. 1977, Anderson and Mozer 1981, Ballard 1986. The
novel aspect to the Pellionisz-Llinas approach is their generalization of the abstract
formalism from vector to tensor analysis and their conception of representation .
2. Lest I be misunderstood on the matter , I should emphasize that the computer
display of the cartoon organism is in no way intended to be a contribution to
computer science. It was merely an (auto)didactic device that made the problem
simple enough so that we could begin to understand the fundamental principles
Pellionisz and Llinas were talking about . Obviously , far more sophisticated robotic
devices have been created in AI labs. However , if the Pellionisz-Llinas approach is
basically right , then it may be that it holds implications for robotics.
3. As Michael Arbib has pointed out to me, there are similarities between the phase
space sandwich model and the model of distributed motor control in the superior
colliculus devised by Pitts and McCulloch in 1947. (See also Arbib 1981.) In the
Churchland model what plays the central role are the notions of phase spacesand
the coordinate transformation of .positions in one phase space to positions in a
different phase space. Hence, the relative deformationof topographically mapped
surfaces is critical in effecting the appropriate transformation . But certainly Pitts
and McCulloch did make the parallel and distributed nature of nervous systems
and the topographic mapping of collicular layers the key elements in their seminal
model .
4. For a discussion critical of the tensor network theory , see Arbib and Amari 1985.
See Pellionisz and Llinas (1985b) for their reply .
5. On this theme, see also Piaget 1971.
6. This example was suggested to me by Ray Jackendoff.
7. McCollum , Pellionisz , and Llinas (1983) have proposed a tensorial hypothesis to
show how the visual system can solve the computational problem . A difficulty
with this germinal hypothesis is that it presents the computation as handled en-
tirely
at
the
retinal
mammals
the
computation
is
significant
8 .
In
This
"
of
global
recent
by
the
color
retina
from
Land
or
whether
problem
also
McClelland
the
Jerry
of
value
Fodor
it
can
be
ni
Oi
11
is
is
the
of
system
external
)
threshold
Other
theorists
Arbib
( 1981
Freeman
the
weight
input
from
. "
( 1975
1982
) ,
the
( Hinton
whose
,
) ,
Gunter
( 1983
indicate
,
done
solved
Rumelhart
and
systematicity
higher
that
whether
up
rather
Hinton
in
the
it
simply
1986
imported
is
1/ 2
'
LL
defined
as
- - 5 -5 '
I]
11
remains
using
the
from
another
theory
is
"
Lil
i ~ ;
( synaptic
is
al
Nevertheless
where
et
constancy
See
energy
data
for
computational
expressing
from
The
whereas
required
approach
conversation
borrowed
10
the
network
way
,
is
executed
that
tensor
9 .
level
cortex
489
to
jth
and
work
Valentino
the
to
ith
the
ith
Sejnowski
ideally
( 1982
should
) ,
unit
unit
,
,
Wij
Si
1983a
Braitenberg
Palm
-s ,
I'
Christoph
is
the
boolean
: 449
have
( 1984
and
is
of
truth
the
connection
value
( 0
or
1 )
and
) .
been
) ,
strength
discussed
Gerald
von
include
Edelman
der
Malsburg
Michael
( 1978
( 1973
) ,
Walter
,
1981
) .
Bibliography
, and J. M . Walker
Allman , John (1977). Evolution of the visual system in early primates . Progressin
Psychobiologyand PhysiologicalPsychology7:1- 53.
Allman , John (1982). Reconstructing the evolution of the brain in primates through the
use of comparative neurophysiological and neuroanatomical data. In Primatebrain
evolution, ed. E. Armstrong and D. Falk, 13- 28. New York : Plenum .
Allman , John , and Jon Kaas (1971) . Representation
adjoining cortex of the owl monkey (Aotus trivirgatus). Brain Research35:89- 106.
Altman , Joseph (1972). Postnatal development of the cerebellar cortex in the rat. Journal
of ComparativeNeurology 145:353- 397.
Anderson , JamesA ., and Geoffrey E. Hinton (1981). Models of information processing
in the brain . In G . E. Hinton
of physiology: The nervous systemII , ed. v' . B. Brooks, 1449- 1480. Bethesda, Md .:
American Physiological Society.
Arbib , Michael (1982). Modelling neural mechanisms of visuomotor coordination in
Frog and Toad . In Competition and cooperation in neural nets, ed . S. Amari
and M .
transformations
in the brain
492
Bibliography
Aserinsky , E., and N . Kleitman (1953). Regularly occurring periods of eye motility , and
concomitant phenomena during sleep. Science118:273- 274.
Atkinson , R. C., and R. M . Shiffrin (1968). Human memory : A proposed system and its
control processes. In The psychologyof learning and motivation, vol . 2, ed. K . W .
Spence and J. T. Spence. New York : Academic Press.
Auburtin , S. A . E. (1861). Reprise de la discussion sur la forme et Ie volume du cerveau.
Bulletin du societed'anthropologiede Paris 2:209- 220. (Excerpted and translated in E.
Clarke and C. D . O'Malley (1968), 492- 494.)
AyTer, A . J., ed. (1959). Logical positivism. New York : Free Press.
Ballard, D . H . (1986). Cortical ,connections and parallel processing: Structure and function . Behavioral
and Brain
Sciences 9 / 1 .
Bantle, J. A ., and W . E. Hahn (1976). Complexity and characterization of polyadenylated RNA in the mouse brain . Cell 8:139- 150.
Barker, C. F., and R. E. Billingham (1977). Immunologically privileged sites. Advances
in Immunology 23 :1- 54 .
Barlow , Horace B. (1972). Single units and sensations: A neuron doctrine for perceptual
physiology ? Perception1:371- 394.
Barr, Murray L . (1974). Thehuman nervoussystem: An anatomicalvielvpoint. 2nd ed. New
York : Harper and Row .
Bartholow , R. (1874). Experimental investigation into the functions of the human brain .
American Journal of Medical Sciences 67:305- 313 .
Bell, Charles (1823). Second paper on the nerves of the orbit . PhilosophicalTransactionso,f
the Royal Society 113:289- 307 .
Bellugi, Ursula , Howard Poizner, and Edward S. Klima (forthcoming ). vVhatthe hands
revealabout the brain. Cambridge , Mass.: MIT Press.
Benzer , Seymour
units
heredity, ed. W . McElroy and B. Glass, 70- 93. Baltimore : The Johns Hopkins Uni versity Press .
during
classical conditioning
Berkeley, George. The Worksof GeorgeBerkeley. (1948- 1957). 9 vols . Ed. A . A . Luce and
T . E. Jessop . London : Nelson .
Berker, Ennis A . (1985). Principles of brain function in neuropsychological development of hydrocephalics . Doctoral dissertation , University of Michigan , Ann Arbor .
Berlin , Charles I . (1977) . Hemisphere
eds . (1977), 303- 323 .
as)'mmetry
in auditory
tasks . In S. Hamad et al .,
Berlucchi, Giovanni , and James M . Sprague (1981). The cerebral cortex in visual learning and memory', and in interhemispheric transfer in the cat. In F. O. Schmitt ,
F . G . Worden , G . Adelman , and S. G . Dennis , eds . (1981), 415- 440 .
Bianchi , Leonardo (1895). The functions of the frontal lobes . (Translated
from the
of Oklahoma
Press .
Bibliography 493
Bisiach, Edoardo, Erminia Capitani, ClaudioLuzzatti, and DanielaPerani(1981
). Brain
and consciousrepresentationof outside reality. Neuropsychologia
19:543- 551.
Bisiach, E., and C. Luzzatti (1978). Unilateralneglectof representationalspace. Cortex
14:129- 133.
Bisiach, E., D. Perani, C. Papagno, and A. Berti (unpublished). Unawarenessof diseasefollowing lesionsof the right hemisphere: Anosognosiafor hemiplegiaand
anosognosiafor hemianopia.
Bjorklund, Anders, and Ulf Stenevi(1977). Reformationof severedhippocampalcholinergic pathway in the adult rat by transplantedseptalneurons. Cell TissueResearch185:289- 302.
Bjorklund, Anders, and Vlf Stenevi(1984). Intracerebralneural implants: Neuronal
replacementand reconstructionof damagedcircuitries. AnnualRevielvof Neurosci
ence7, 279- 308.
Blakemore
, Colin (1977). Mechanicsof the mind. Cambridge: CambridgeUniversity
Press.
Blakemore
, Colin, and SusanGreenfield, eds. (in press). Mind matters
. Oxford: Basil
Blackwell.
Blanks, R. H. I., I. S. Curthoys, and C. H. Markham(1975). Planarrelationshipsof the
semicircularcanalsin man. ActaOtolaryngolica
80:185- 196.
Block, Ned (1978). Troubleswith functionalism. In Perception
andcognition
. Issuesin the
foundationof psychology
, 261- 325. (MinnesotaStudiesin the Philosophyof Science
9.) Minneapolis: University of MinnesotaPress. (Reprintedin N. Block, ed. (1981).
Readings
in thephilosophy
of psychology
, 268- 305. Cambridge, Mass.: Harvard University Press.)
Bloom, F. E. (1980). Neuropharmacologyand the adaptiveregulationof receptorsensitivity: An overview. Neurosciences
Research
Pro- ramBulletin18/3:429- 435.
Bodian, R. (1967). Neurons, circuits, and neuroglia. In Theneurosciences
: A studyprogram, ed. G. C. Quartan, T. Melnechuck, and F. O. Schmitt, 6- 24. New York:
RockefellerUniversity Press.
BogenJ. E., E. D. Fisher, and P. J. Vogel (1965). Cerebralcommissurotomy
: A second
casereport. Journalof theAmericanMedicalAssociation
194:1328- 1329.
Bonin, G. von, ed. (1960). Somepaperson thecerebral
cortex.Springfield, Ill .: CharlesC.
Thomas.
80010s
, George5., and RichardC. Jeffrey(1974
). Computability
andlogic. Cambridge:
CambridgeUniversity
- Press.
Bower, J., and R. Llinas (1983). Simultaneoussamplingof the responsesof multiple,
closelyadjacent, Purkinje cells respondingto climbing fiber activation. Society
for
Neuroscience
Abstracts9:607.
Bradshaw,J. L., and N. C. Nettleton (1981). The natureof hemisphericspecialization
.
Behavioral
andBrainSciences
4/1:51- 63.
Braestrup, Claus, and Mogens Nielsen (1980
). Multiple benzodiazepinereceptors.
Trendsin Neurosciences
3:301- 303.
Braitenberg
, Valentino(1984). Vehicles
. Cambridge, Mass.: MIT Press.
Braitenberg
, V., and N. Onesto (1961). The cerebellarcortexasa timing organ. Discussion of an hypothesis. Proceedings
of the1stAnnualConference
onMedicalCybernetics
,
1- 19. Naples: Giannini.
Brazier, Mary A. B. (1982). The problem of neuromuscularaction: Two 17th century
Dutchmen. In Historicalaspects
of theneurosL
'iences
, ed. F. C. Roseand W. F. Bynum,
13- 22. New York: Raven.
Brazier, Mary A. B. (1984). A historyof neurophysiology
in the17thand18thcenturies
. New
York: Raven.
494
Bibliography
tion . In Attention: Contempprarytheory and analysis, ed. D . Mostofsky , 51- 60. New
York : Appleton -Century -Crofts .
Broca, Pierre-Paul (1861
_). Remarques sur Ie siege de la faculte du langage articuJe;
suivies d'une observation d'aphemie (perte de la parole). Bulletin de la societe
anatomique de Paris 36:330- 357 . (Translated as " Remarks on the seat of the faculty of
Broca, Pierre-Paul (1863). Localisation des fonctions cerebrales: Siege du langage articule . Bulletin de LaSocieted'anthropologie. Paris 4:200- 202.
Brodmann , Korbinian (1909). VergleichendeLokalisationlehreder Grosshirnrindein ihren
Prinzipien dargestelltauf Grund des Zellenbaues
. Leipzig : J. A . Barth.
Brookshire, R. H ., and L . E. Nicholas (1984). Comprehension of directly and indirectly
stated main ideas and details in discourse by brain -damaged and non-brain damaged listeners . Brain and Language21:21- 36.
Brown , C. M . (1984). Computer vision and natural constraints . Science224:1299- 1305.
Bryden, M . P. (1982).. Laterality: Functional asymmetry in the intact brain. Ne\l\' York :
Academic
Press .
York : Raven .
. San Francisco
: W . H . Freeman
: W . H . Freeman
Bunge, Richard P. (1968). Glial cells and the central myelin sheath. PhysiologicalReview
48 : 197 - 251 .
Burklund , C. W ., and Aaron Smith (1977). Language and the cerebral hemispheres.
Neurology 27:627- 633 .
Butterfield , Herbert (1968). The origins of modern science1300- 1800. Toronto : Clarke,
Irwin
and
Co .
Campbell , Keith (1983). Abstract particulars and the philosophy of mind . Australasian
Journal of Philosophy61:129- 141.
Campion , J., R. Latto , and Y. M . Smith (1983). Is blindsight an effect of scattered light ,
spared cortex, and near-threshold vision ? Behavioraland Brain Sciences6:423- 486.
Carew, T. J., E. T. Walters, and E. R. Kandel (1981). Classical conditioning in a simple
withdrawal
Carnap , Rudolf
Bibliography
495
lated by R. A . George (1967). The logical structure of the zvorld. London : Routledge
and Kegan Paul .)
Carnap, Rudolf (1932). Uberwindung der Metaphysik durch logische Analyse der
Sprache. Erkenntnis 2:219- 241. (Translated by A . Pap as The elimination of
metaphysics through logical analysis of language. Reprinted in A . J. Ayer , ed.
(1959). Logicalpositivism, 60- 81. New York : Free Press.)
Carnap, Rudolf (1950). Empiricism , semantics and ontology . Revue internationale de
philosophie4:20- 40. (Reprinted (1956) in Meaning and necessity
, 2nd ed., 205- 221.
Chicago: Chicago University Press.)
Chihara, C. S., and Jerry Fodor (1965). Operationalism and ordinary language: A
critique of Wittgenstein . American PhilosophicalQuarterly 2:281- 295.
Chomsky , Noam (1965). Aspectsof the theory of syntax. Cambridge, Mass.: MIT Press.
Christensen , B ., H . Krausz , R . Perez -Polo , and W . D . Willis , Jr. (1980) . Communication
among neurons and neuroscientists
ed . H . M . Pinsker
and
W . D . Willis
York : Raven .
Philosophy77/4:185- 207.
Churchland
Nous
Churchland
processing .
14 : 147 - 170 .
referral of experiences
and its relevance to the mind -body problem . Philosophyof Science48:165- 181.
Churchland , Patricia Smith (1981b). The timing of sensations: Reply to Libet . Philosophy
of Science 48 :492- 497 .
, Patricia Smith
of a concept .
Churchland
79 :214 - 236 .
259 .
Churchland
materialism
attitudes .
of brain
496
Bibliography
Clarke, Edwin , and (i:. D . O'Malley (1968). The human brain and spinal cord: A historical
study illustrated by uJritingsfrom antiquit!/ to the tuJentiethcentury. Berkeley and Los
Angeles : University of California Press.
Clement , John (1982). Students' preconceptions in introductory mechanics. American
Journal of Physics50:66- 71.
Clemes, 5., and W . Dement (1967). The effect of REM sleep deprivation on psychologi cal functioning . Journal of Nervousand Mental Disorders144:485- 491.
Cohen, N . J., and S. Corkin (1981). The amnesic patient . H . M .: Learning and retention
of cognitive skill . Societyfor NeuroscienceAbstracts7:517- 518.
Constantine -Paton , Martha (1981) . Induced ocular -dominance zones in tectal cortex . In
F. O . Schmitt , F . G . Worden , G . Adelman , and 5 . G . Dennis , eds . (1981), 47- 67 .
Constantine -Paton, Martha , and Margaret I . Law (1982). The development of maps and
stripes in the brain . ScientificAmerican 247/6:62- 70.
Cooper, J. R., F. E. Bloom, and R. H . Roth (1982). Thebiochemicalbasisof neuropharmacol
ogy. 2nd ed. New York : Oxford University Press.
Corkin , Suzanne (1968). Acquisition of motor skill after bilateral medial temporal -lobe
excision. Neuropsychologia6:255- 265.
Comsweet , Tom N . (1970) . Visual perception . New York : Academic
Press .
Cowan , W . Maxwell (1979). The development of the brain . Scientific American 241/
3 : 112 - 133 .
Cowey , A . (1981). Why are there so many visual areas? In F. O. Schmitt, F. G. Worden ,
G . Adelman , and S. G . Dennis , eds . (1981), 395- 413 .
Coyle , J. T ., D . L . Price , and M . R . Delong (1983) . Alzheimer ' s disease : A disorder
of
for
of the thalamic
reticular
: 111 - 114 .
Critchley , Macdonald (1962). Speech and speech-loss in relation to the duality of the
brain . In Interhemisphericrelationsand cerebraldominance
, ed. V . B. Mountcastle , 208213 . Baltimore : The Johns Hopkins
University
Press .
Critchley , MacDonald (1979). The divine banquetof the brain. New York : Raven.
Crow , T. J., A . J. Cross, Eve C. Johnstone, and F. Owen (1982). Two syndromes in
schizophrenia and their pathogenesis. In F. A . Henn and H . A . Nasrallah, eds.
(1982), 196- 234 .
H . A . Nasrallah
, eds . ( 1982 ) , 40 - 60 .
Bibliography
497
Davidson , Donald (1970). Mental events. In Experienceand theory, ed. L . Foster and
J. W . S\"-"anson, 79- 101. Amherst , Mass.: University of Massachusetts Press.
Davis, J. M ., and D . L . Garver (1978). Neuroleptics : Clinical use in psychiatry . In
Handbookof psychopharmacology
, vol . 10, ed. L . L . Iverson , S. D . Iverson , and S. H .
Snyder, 129- 164. New York : Plenum .
Davis, William J. (1980). Neurophilosophical reflections on central nervous pattern
generators . Behavioral and Brain Sciences 3:543- 544 .
Dawkins , Richard , and John Krebs (1978) . Animal
signals : Information
or manipula -
: Basil
Blackwell
Deiters , Otto F . K . (1865). Untersuchun :(,en uber Gehirn und Ruckenmark des Menschen und
der Siiugethiere
. (Posthumous .) Braunschweig : Vieweg . (Excerpted and translated
in E. Clarke and C . D . O 'Malley
callosum
: University
of Minnesota
Press . (Reprinted
in D . C . Dennett (1978b) .)
Press
and
Harvester
Press .
, Holland
: D . Reidel .
1982:32- 39 .)
Dennett , Daniel C. (1983). Intentional systems in cognitive ethology : The " Panglossian
Paradigm" defended . Behavioraland Brain Sciences6:343- 390.
Dennett , Daniel C. (1984a). Cognitive wheels: The frame problem of artificial intelli gence. In Minds, machines
, and evolution, ed. C. Hookway , 129- 151. Cambridge :
Cambridge University Press.
Dennett , Daniel
. : MIT
Press
498
Bibliography
Descartes, Rene (1649). Les Passionsde l'ame. English translation in E. S. Haldane and
G. R. T . Ross (1911). The philosophicalworks of Descartes
. 2 vols. Reprinted (1968).
Cambridge : Cambridge University Press.
Descartes, Rene (1664). Traite de l'homme. Paris: Angot . French text with English translation in Thomas Steele Hall (1972). Treatiseon man. Cambridge, Mass.: Harvard
University Press.
Desimone, R., T. D . Albright , C. G. Gross, and C. Bruce (1984). Stimulus -selective
properties
4 : 2051 - 2062
of inferior
temporal
neurons
Desmedt, John E., Julien Debecker, and Donald Robertson (1979). Serial perceptual
processing and the neural basis of changes in event-related potentials components
and slow potential shifts . In J. E. Desmedt, ed. (1979), 53- 79.
Deutsch , Diana (1984) . Memory
for nonverbal
information
and physiological studies. In L. R. Squire and N . Butters, eds. (1984), 45- 54.
DeVoogd, T. J., and F. Nottebohm (1981). Sex differences in dendritic morphology of a
song control nucleus in the canary: A quantitative Golgi study . Journalof Comparative Neurology 196:309- 316 .
Dewan , E. M . (1976). Consciousness as an emergent causal agent in the context of
control system theory . In Consciousness and the brain , ed . G . Globus , G . Maxwell
and
I . Savodnik
York
: Plenum
correlates
of pattern
recognition . In International
Dreyfus , Hubert L . (1979). What computerscan't do. 2nd ed. New York : Harper and Row .
Dreyfus , Hubert
to knowledge
representation
: AI at an
impasse. In Mind design, ed. J. Haugeland , 161- 204. Cambridge, Mass.: MIT Press.
Du Bois-Reymond , Emil (1849). UntersuchungenuberthierischeElektricitiit. Vol . 2. Berlin :
Reimer. (Excerpted and translated in E. Clark and C. D. O'Malley (1968), 197.)
Duhem , Pierre (1914). La theoriephysique: Son object, sa structure. Paris: Riviere et Cie.
(Translated (1953) by Philip Weiner as Theaim and structure of physicaltheory. Princeton , N .J.: Princeton University Press. Reprinted (1962). New York : Atheneum .)
Duncan-Johnson, C. C., and E. Donchin (1977). On quantifying surprise : The variation
of event-related potentials with subjective proba.bility . Psychophysiology
14:456- 467.
Eccles, John C. (1977). Part II of Theselfand its brain (Popper and Eccles1977), 225- 406.
Eccles , John C ., M . Ito , and J. Szentagothai
Press .
Bibliography 499
Edelman, GeraldM. (1978). Group selectionand phasicreentrantsignalling: A theory
of higher brain function. In Themindfulbrain, G. M. Edelmanand V. B. Mountcastle, 51- 100. Cambridge, Mass.: MIT Press.
Edelman, Robert (1984). The clinicians's guide to the theory and practiceof NMR
scanning. Discussions
in Neurosciences
1/1:9- 48.
Ehrhardt, A. A., and S. W. Baker(1974). Fetalandrogens, humanCNSdifferentiation,
and behavior sex differences. In Sexdifferences
in behavior
, eds. R. C. Friedman,
R. M. Richart, and R. L. van de Wiele, 33- 51. New York: Wiley.
Einstein, Albert (1949). Autobiographicalnotes. In AlbertEinstein
: Philosopher
-Scientist
,
ed. P. Schilpp, 3- 95. Evanston, Ill .: The Library of Living Philosophers
.
En\ , Berent(1983). In defenseof the identity theory. Journalof Philosophy
80:279- 298.
Fahlman, Scott (1979). NETL: A systemfor representing
and usingreal-u70rldknowledge
.
Cambridge, Mass.: MIT Press.
Fahlman, Scott (1981). Representingimplicit knowledge. In G. E. Hinton and J. A.
Anderson, eds. (1981), 145- 159.
Feder, H. H. (1984). Hormones and sexual behavior. Annual Reviewof Psychology
35:165- 200.
Feindel, W., R. S. J. Frackowiak, D. Gadian, P. L. Magistretti, and M. R. Zalutsky, eds.
(1985). Brain metabolismand imaging. Discussions
in Neurosciences
2/2 (special
issue).
Feldman, J. A. (1985). Connectionistmodelsand their applications:Introduction. CognitiveScience
9:1- 2 (specialissue).
Feldman, J. A., and D. H. Ballard(1982). Connectionistmodelsand their properties.
CognitiveScience
6:205- 254.
Feldman, RobertS., and Linda F. Quenzer(1984
). Fundamentals
ofneuropsychopharmacol
ogy. Sunderland, Mass.: Sinauer.
Ferraro, P., A. Guidotti, B. Conti-Tronconi, and E. Costa (1984
). A brain octadecaneuropeptidegeneratedby tryptic digestion of OBI (DiazepamBinding Inhibitor) functions as a proconflict ligand of benzodiazepinerecognition sites.
Neuropharmacology
23:1359.
Ferrier, David (1876). Thefunctionsof thebrain. London: Smith, Elder.
Ferrier, David (1890). TheCroonianlectures
oncerebral
localisation
. London: Smith, Elder.
Feyerabend
, PaulK. (1962). Explanation, reduction, and empiricism. In Minnesota
studies in the philosophy
of science
, vol. 3, ed. H. Feigl and G. Maxwell, 28- 97. Minneapolis: University of MinnesotaPress. (Reprintedin P. K. Feyerabend(1981).
Philosophical
papers
, vol. I , 44- 96. Cambridge: CambridgeUniversity Press.)
Feyerabend
; Paul K. (1963a
). How to be a good empiricist: A plea for tolerancein
mattersepistemological
. In Philosophy
, theDelazvare
seminar
, vol. 2, ed. B.
~ of science
Baumrin, 3- 39. New York: Interscience
. (Reprintedin H. Morick, ed. (1972), 164193.)
Feyerabend
, Paul K. (1963b
). Materialismand the mind-body problem. TheRevieuJ
of
Metaphysics
17:49- 66. (Reprintedin C. V. Borst, ed. (1970
). Themind/brainidentity
theory
- , 142- 156. London: Macmillan.),
Feyerabend
, PaulK. (1963c
). Mental eventsand the brain. Journalof Philosophy
60:295296.
Feyerabend
, Paul K. (1981). Philosophical
papers
. Vols. 1 and 2. Cambridge: Cambridge
University Press.
Field, Hartry (1977
) . Logic, meaning, andconceptualrole. Journal
ofPhilosophy
74:379- 408.
Field, Hartry (1978). Mental representation
. Erkenntnis
13:9- 61. (Reprintedin N. Block,
ed. (1981). Readings
in thephilosophy
ofpsychology
, vol. 2, 78- 114. Cambridge,Mass.:
Harvard University Press.)
500 Bibliography
Fine, Michael , Donald A . Keefer, and George R. Leichnetz (1982). Testosterone uptake
in the brainstem of a sound -producing fish . Science215/4537:1265- 1267.
Flanagan, Owen J" Jr. (1984). The sciel1ce
of the 111ind
. Cambridge, Mass.: MIT Press.
Flourens, Pierre (1824). Recherches experimentales sur les proprietes et les fonctions
du systeme nerveux dans les animaux vertebres. (Translated as " Investigation of
the properties and the functions of the various parts which compose the cerebral
mass" in G. von Bonin , ed. (1960), 3- 21.)
Fodor, Jerry A . (1974). Special sciences. Synthese28:77- 115. (Reprinted in J. A . Fodor
(1981).)
Fodor, Jerry A . (1975). The languageof thou,ght. New York : Crowell . (Paperback edition
(1979). Cambridge , Mass.: Harvard University Press.)
Fodor, Jerry A . (1981). Representations
. Cambridge, Mass.: MIT Press.
Fodor, Jerry A . (1985). Fodor's guide to mental representation : The intelligent auntie ' s
vade-mecum . Mind 94:76- 100.
Forel, August -Henri (1887). Einige hirnanatomische Betrachtungen und Ergebniss~.
Arch. Psychiat. NervKrankh. 18:162- 198. (Excerpted and translated in E. Clarke and
C. D . O'Malley (1968), 105- 109.)
Foss, Jeffrey (1986). Abstract solutions vs. neurobiologically plausible problems . (Commentary on Ballard 1986.) BeJlavioraland Brain Sciences9.
Foulkes, D . (1966). The psychologyof sleep. New York : Scribner' s.
Fox, Jeffrey I... (1984). PET scan controversy aired . Science224:143- 144.
Freed, W . J. (1983). Functional brain tissue transplantation : Reversal of lesion-induced
rotation by intraventricular substantia nigra and adrenal medulla grafts, with a
note on intracranial retinal grafts . BioLogicaL
Psychiatry 18/11:1205- 1267.
Freeman, Walter (1975). Mass action in the nervoussystem. New York : Academic Press.
Freeman, Walter (1979). Nonlinear dynamics of paleocortex manifested in the olfactory
EEG. BiologicalC~/berrleti(~s 3:21- 37.
Frege, Gottlob (1952). Translationsfrom the philosophical'lvritings of Gottlob Frege, trans .
and ed. P. Geach and M . Black. Oxford : Basil Blackwell .
Fritsch, G., and E. Hitzig (1870). Uber die elektrische Erregvarkeit des Grosshirns .
Archiv fur Anatomie, PhysioLogieu1'ld Wissenschaftliche
Medizill . (Translated as " On
the electrical excitability of the cerebrum" in G. von Bonin, ed. (1960), 73- 96.)
Furshpan , E. J., and D . D . Potter (1959). Transmission at the giant motor synapses of
the crayfish . Journal of Physiology145:298- 325.
Gainotti , Guido , Carlo Caltagirone , and Gabriele Miceli (1983). Selective impairment of
semantic-lexical discrimination in right -brain -damaged patients . In E. Perecman,
ed. (1983), 149- 167.
Galaburda, Albert M ., and Marek -Marsel Mesulam (1983). Neuroanatomical aspects of
cerebral localization . In A . Kertesz, ed. (1983b), 21- 61.
Galambos, Robert, and Steven A . Hillyard , eds. (1981). Electrophysiological
approaches
to human cognitive processing
. (Based on a work session of the Neurosciences Research Program, 1981.) Neurosciences
ResearchProgram Bulleti1'l 20:141- 265.
Galen. Galen on anatomicalprocedures
. Tral1sLation
of the surIJivingbooks{{Jitli il1troductioll
and notes. (1956). C. Singer. London : Oxford University Press.
Gall, Franz Josef (1812). Anatomieet ph~/siologiedll systemenerveuxengenl~ral, ('t du L'erl'eau
en particulier. Vol . 1. Paris: Schoell. (Excerpted and translated in E. Clarke and
C. D . O'Malley (1968), 476- 480.)
Gardner , Howard , Hiram H . Brownell , Wendy Wapner, and Diane Michelow (1983).
Missing the point : The role of the right hemisphere in the processing of complex
linguistic material . In E. Perecman, ed. (1983), 169- 191.
Gazzaniga, Michael S (1970). The bisectedbrain. New York : Appleton -Centurv -Crofts .
Bibliography
501
502
Bibliography
Bibliography 503
Hebb, D. O . (1939). Intelligence in man after large removals of cerebral tissue: Report of
four left frontal lobe cases. Journal of GeneralPsychology21:73- 87.
Hebb, D. 0 ., and W . Penfield (1940). Human behavior after extensive bilateral remov als from the frontal lobes. Archives of- Neurolo~~
!l.:t/ and Pstjchiatn
; 44:421- 438.
vHecaen,
v
Henri , and Martin L . Albert , eds. (1978). Human neuropsychology
. New York :
Wiley .
Heilman , Kenneth M . (1979). Neglect and related disorders . In K . M . Heilman and E.
Valenstein , eds. (1979), 268- 307.
Heilman , Kenneth M ., Gregory Howell , Edward Valenstein, and Leslie Rothi (1980).
Mirror -reading and writing in association with right -left spatial disorientation .
Journal of Neurology, Neurosurgery, and Psychiatry43:774- 780.
Heilman , Kenneth M ., and Edward Valenstein , eds. (1979). Clinical neuropsychology
.
New York : Oxford University Press.
Heilman , Kenneth M ., Robert R. Watson, Edward Valenstein, and Antonio R. Oamasio
(1983). Localization of lesions in neglect. In Andrew Kertesz, ed. (1983b), 471- 492.
Heimer , Lennart (1983). The human brain and spinal cord: Functional neuroanatomyand
dissectionguide. New York : Springer-Verlag .
Heimer , Lennart , and Martine J. Robards, eds. (1981). Neuroanatomicaltract-tracing
methods. New York : Plenum .
Held , Hans (1897). Beitrage zur Structur der Nervenzellen und ihrer Fortsatze. Zweite
Abhandlung . Arch. Anat. Physiol. (Anat. Abt.), 204- 294. (Excerpted and translated
in E. Clarke and C. D . O'Malley (1968), 119- 120.)
Helmholtz , Hermann L . F. von (1850). Vorlaufiger Bericht fiber die Fortpflanzungsge schwindigkeit der Nervenreizung . Archivfur Anatomie, Physiologieund wissenschaft
liche Medizin, 71- 73. (Translated in W . Dennis (1948). Readingsin psychology
, 197198. New York : Appleton -Century -Crofts .)
Hempel , Carl (1945). A definition of " degree of confirmation ." Philosophyof Science
12:98- 115.
Hempel , Carl (1965). Aspectsof scientificexplanation. New York : Free Press.
Henn , Fritz , A ., and Henry A . Nasrallah , eds. (1982). Schizophreniaas a brain disease
.
New York : Oxford University Press.
Heritage, Aileen S., and Walter Stumpf (1980). Brain catecholamine neurons are target
sites for sex hormones . Science207,/4437:1377- 1379.
Hesse, Mary (1970). Is there an independent observation language? In The nature and
function of scientific theories, ed. R. Colodny , 36- 77. Pittsburgh : University of Pittsburgh Press.
Heuser, J. E., and T. S. Reese (1979). Synaptic vesicle exocytosis captured by quick freezing . In F. O. Schmitt and F. G. Worden , eds. (1979), 573- 600.
Hillyard , S. A ., G. V . Simpson, D . L . Woods, S. van Voorhis , and T. Munte (1984).
Event-related brain potentials and selective attention to different modalities . In
Neural integrationat basicand corticallevels: IBRa conference
, ed. F. Reinoso, 395- 414.
New York : Raven.
Hillyard , S. A ., and M . Kutas (1983). Electrophysiology of cognitive processing. Annual
review of psychology34:33- 61.
Hinton , Geoffrey E. (1981). Implementing semantic networks in parallel hardware . In
G. E. Hinton and J. A . Anderson , eds. (1981), 161- 187.
Hinton , Geoffrey E., and James A . Anderson , eds. (1981). Parallel modelsof associative
memory. Hillsdale , N .J.: Erlbaum .
Hinton , Geoffrey E., And Terrence J. Sejnowski (1983a). Optimal perceptual inference.
Proceedings
of the IEEE Conference
on computervision and pattern recognition(Washington, D .C., June 1983), 448- 453. Silver Spring, Md .: IEEE Computer Society Press.
504
Bibliography
Hinton , Geoffrey E., and Terrence J. Sejnowski (1983b). Analyzing cooperative compu tation . In Proceedingsof the 5th annual conferenceof the Cognitive ScienceSociety,
Rochester , N . Y .
Hinton , Geoffrey E., and Terrence J. Sejnowski (1986). Learning and relearning in
Boltzmann
machines . In D . E . Rumelhart
Hirano , Asao, and Herbert M . Dembitzer (1967). A structural analysis of the myelin
sheath in the central nervous system. Journal of Cell Biology 34:555- 567.
His , Wilhelm (1888). Uber die embryonale Entwickelung der Nervenbahnen . Anat.
Anz . 3:499- 505. (Excerpted and translated in E. Clarke and C. D. O'Malley (1968),
103- 104.)
Hitzig , G., and G. T. Fritsch (1870) Uber die elektrische Erregbarkeit des Grosshirns .
Archiv fur anatomie, Physiologieund wissenschaftliche
Medizin, 300- 332. (Translated as
" On the electrical excitability of the cerebrum" in G. von Bonin , ed. (1960), 73- 96.)
Hodgkin , A . L ., and A . F. Huxley (1952). A quantitative description of membrane
current and its application to conduction and excitation in nerve. Journalof Physiology 117:500- 544. (Reprinted in C. D . Barnes and C. Kircher , eds. (1968). Readingsin
neurophysiology
, 205- 249. New York : Wiley .)
Holtzman , Jeffrey D ., and Michael S. Gazzaniga (1982). Dual task interactions due
exclusively to limits in processing resources. Science218:1325- 1327.
Hooker , C. A . (1975). Systematic philosophy and meta-philosophy of science: Empiri cism, Popperianism , and realism . Synthese32:177- 231.
Hooker , C. A . (1981). Towards a general theory of reduction . Part I : Historical and
scientific setting . Part II : Identity in reduction . Part III : Cross-categorial reduction .
Dialogue20:38- 59, 201- 236, 496- 529.
Hopfield , J. J., D . I . Feinstein, and R. G. Palmer (1983). " Unlearning " has a stabilizing
effect
in collective
memories
. Nature
Horgan , Terrence, and JamesWoodward (1985). Folk psychology is here to stay. PhilosophicalReview44:197- 226.
Horn , G. (1983). Information storage in the brain : A study of imprinting in the domestic
chick . In Advancesin vertebrateneuroethology
: Proceedingsof a NATO AdvancedStudy
Institute, ed. I .-P. Ewart , R. Capranica, and D . J. Ingle , 511- 541. New York :
Plenum
Hrdy , Sarah Blaffer (1981). The woman that never evolved. Cambridge, Mass.: Harvard
University Press.
Hubel , D . H ., and T. N . Wiesel (1965). Receptive fields and functional architecture in
two non -striate visual areas (18 and 19) of the cat. Journalof Neurophysiology28:229289 .
(1972) . Laminar
and columnar
distribution
of
(1977) . Functional
architecture
of macaque visual
Bibliograph )T
505
506
Bibliography
Kaas, J. H ., R. J. Nelson , M . Sur, and M . M . Merzenich (1979). Multiple representations of the body within the primary somatosensory cortex of primates . Science
204 : 521 - 523 .
Kanaseki, T., and J. M . Sprague (1974). Anatomical organization of pretectal nuclei and
tectallaminae in the cat. Journal of ComparativeNeurology 158:319- 337.
Kandel , Eric R. (1976). Cellular basisof behavior: An introduction to behavioralneurobiology.
San Fransisco
: W . H . Freeman
Kandel , Eric R . (1981c) . Somatic sensory system III : Central representation of touch . In
E . R . Kandel and J. H . Schwartz , eds . (1981), 184- 198.
Kandel , Eric R., and James H . Schwartz , eds . (1981). Principles of neural science. New
York , Amsterdam
Kant , Immanuel
, Oxford
: Elsevier
/ North
- Holland
(1781) . Kritik der reinen Vernun .ft . 2nd ed . (1787) . Translated and edited
: Macmillan
Kelley , Darcy B. (1980). Auditory and vocal nuclei in the frog brain concentrate sex
hormones
Kelly , James P. (1981a). Visual system II : Anatomy of the central visual pathways . In
E . R . Kandel and J. H . Schwartz , eds . (1981), 226- 235 .
Kety , S. S. (1978). The biological roots of mental illness: Their ramifications through
cerebral metabolism , synaptic activity , genetics , and the environment
ture
. Harvey Lec-
71 : 1 - 22 .
C . Thomas
507
Bibliography
and
hemispheric
Friedland
specialization
Kinsbourne
present
41 - 49
Marcel
and
Kinsbourne
Marcel
Marcel
Kinsbourne
of
Kinsbourne
New
York
Kirkpatrick
Philip
Kitcher
Philip
: 335
Koester
- 373
,
,
, eds
- 987
learning
and
and
Kolers
Paul
ogy
Konishi
7 : 289
ior
. Trends
,
) .
of
Jargon
27 : 296
- term
, ed
. D . Deutsch
Vecchi
functional
aphasia
on
- 299
Neuropsy
agnosia
color
memory
processes
and
( 1983
J . A
) .
and
the
. Deutsch
, 257
Optimization
psychology
, reduction
( 1979
am
- 291
by
simulated
? Philosophy
of
Science
, and
functional
psychology
. Phi
for
the
Philosophy
tale
of
) .
The
signs
of
two
Science
33 : 337
sciences
language
Cambridge
- 359
Review
of
Mass
.:
van
Helmholtz
. English
translation
by
F . A
. Welby
.
potential
, P . Rovamo
associative
) .
. In
E . R . Kandel
),
Pekka
105
and
- 143
, K . Bry
. Neuroscience
Lehti6
( 1981
memory
: New
Report
, J . Hyvarinen
memory
, and
) . Report
Q
. H
rats
J. H
. Schwartz
) . Storage
systems
, and
2 : 1065
and
In
L . Vainio
- 1076
( 1977
processing
G .
E .
) .
.
of
Hinton
infor
and
J.
- 776
A .
.
drugs
Sex
and
the
hormones
: Grafts
brain
and
. Freeman
Q
( 1980
. Science
brain
205
/ 4408
development
: 774
.
Science
205
brain
damage
) . Fundamentals
. Science
of
human
217
/ 4557
neurops
: 342
- 344
~/ chology
.
. San
Whishaw
.
Brain
) . Specificity
of
( 1984
) .
Decortication
Research
operation
abolishes
11 / 2 : 123
in
sentence
- 134
place
but
not
cue
recognition
. Cognitive
Psychol
.
( forthcoming
.
, and
in
correct
. Whishaw
. Behavioral
, Masakazu
) . Centrally
Mark
E . Gurney
Neurosciences
Stephen
Psychology
human
significance
- 705
) . Observations
Psychiatry
P .
that
associative
Ian
- 306
, Masakazu
Neurosciences
of
Lehtio
Oja
Ian
( 1975
in
Konishi
Kosslyn
in
: W
Bryan
membrane
) . Report
( 1982
,
Francisco
: Past
. ( 1981
( 1979b
Gina
Bryan
mind
.
growth
localizing
86 : 697
( 1963b
Journal
all
Hermann
neural
( 1979a
: 985
the
) . The
) . Short
: Genes
Bellugi
) .
, Pekka
of
Gina
. British
) . Resting
, Erkki
and
Press
and
: Dover
distributed
Kalata
Body
reduce
- 636
1953
( 1906
in
, Gina
Kolata
: 633
Press
) , 27 - 35
Anderson
and
- 680
to
Ursula
, Teuvo
Kolata
4410
) .
York
, Teuvo
mation
Kolb
L .
principle
Kohonen
: 671
) . Genes
and
, J . ( 1981
In
) . New
Kohonen
( 1975
memory
) . Discussion
University
. ( 1981
R . P .
How
49
( 1984
eds
) .
( 1982
Edward
( 1965
: Academic
. Brain
( 1964
, and
- term
Gelatt
220
( 1982
Koenigsberger
and
( 1963a
Warrington
F . Wood
Short
D .
Science
Harvard
. Weinstein
, Patricia
of
mind
and
- 420
Warrington
Press
( 1980
- 140
Kitcher
Kolb
C .
Patricia
losophy
93
and
. Science
: 134
Kitcher
. In
S .,
: 411
perception
Neurosurgery
: Academic
annealing
on
York
. New
Warrington
form
Elizabeth
,
Marcel
syndrome
limitations
- 175
37
Elizabeth
, and
Neurology
nesic
. E . A
specialization
Elizabeth
and
, ed
, Marcel
Journal
47
and
1 : 27 - 37
155
Hemisphere
visual
Marcel
chologia
Klima
- based
,
neurology
Psychologist
simultaneous
Kinsbourne
Brain
) .
in
: Raven
. R . Rieber
. American
limited
Kitcher
) .
ed
( 1982
understanding
Kinsbourne
York
( 1980
future
. Advances
. New
M
7 : 341
( 1975
- 370
5 / 1 : 20 - 23
) .
Information
synthesized
( 1982
) . Sex
maps
of
differentiation
sensory
of
space
brain
and
images
. Trends
behav
.
representation
in
visual
Cognitive
508
Bibliography
, Mass . : Sinauer
approachto the function of the nervoussystem. 2nd ed. Sunderland , Mass.: Sinauer.
Kuhn , Thomas S. (1962). The structure o.f scientific revolutions. 2nd ed. (1970). Chicago:
University of Chicago Press.
Kuhne , Willy (1862). Uber die peripherischenEndorganeder motorischenNerven. Leipzig :
Engelmann . (Excerpted and translated in E. Clarke and C. D . O'Malley (1968), 7577.)
York
: Raven
P300
as
a measure
of
stimulus
evaluation
time
. Science
La Mettrie , Julien affray de (1748). L 'Hommemachine. (Published anonymously .) Leyden: Luzac. (Translation into English by E. Luzac and G. Smith (1750), London ,
and G. C. Bussey (1912), Chicago.)
Land , Edwin (1977). The retinex theory of color vision . ScientificAmerican237:108- 128.
Land , Edwin H ., David H . Hubel , Margaret S. Livingstone , S. Hollis Perry , and
Michael M . Bums (1983) . Colour -generating interactions across the corpus cal losum
. Nature
Larsen , T . Andreo , and Donald B. CaIne (1982). Recent advances in the study
Parkinson ' s disease . (Review .) Trends in Neurosciences 5/1:10- 12.
of
Lee, K . S., F. Schottler, M . Oliver , and G. Lynch (1980). Brief bursts of high -frequency
stimulation produce two types of structural change in rat hippocampus . Journalo.f
Neurophysiology 44:247- 258 .
Lent , Charles M . (1980) . On neuronal
and Brain
nihilism : Commentary
on Selverston . Behavioral
in writing
194 :337 .
(1972) . Perception
of bilateral
chimeric
Bibliography
509
Lewis , David (1983). Postscript to " Mad pain and Martian pain ." In Philosophicalpapers,
vol . I, 122- 130. New York : Oxford University Press.
Leyton (Grunbaum ), A . S. F., and C. S. Sherrington (1917). Observations on the excitable cortex of the chimpanzee, orangutan , and gorilla . Quarterly Journal of Experimental Physiology11:135- 222. (Reprinted in G. von Bonin, ed. (1960). 283- 396.)
Libet , Benjamin
York : Academic
Press .
Llinas , Rodolfo R. (1975) . The cortex of the cerebellum . Scientift.c American 232/ 1:56- 71.
Llinas , Rodolfo R . (1979). The role of calcium in neuronal function . In F. O . Schmitt and
F . G . Worden , eds . (1979), 555- 571 .
Physiological
~
Society .
-'
electrobiology
of mammalian
central neurons .
Press .
Llinas, Rodolfo R. (1984c). Rebound excitation as the physiological basis for tremor : A
biophysical study of the oscillatory properties of mammalian central neurones in
vitro. In Movement disorders: Tremor, ed. L . J. Findley and R. Capildeo, 165- 182.
London
: Macmillan
Llinas , Rodolfo R. (1984d ). Possible role of tremor in the organization of the nervous
system . In Movement disorders : Tremor , ed . L . J. Findley and R . Capildeo , 475- 477 .
London
: Macmillan
Llinas, Rodolfo R. (forthcoming ). " Mindness " as a functional state of the brain . In Mind
matters , ed . C . Blakemore
and
S . Greenfield
. Oxford
: Basil
Blackwell
Llinas, Rodolfo R., R. Baker, and C. Sotelo (1974). Electronic coupling between neurons
in the cat inferior olive . Journal of Neurophysiology37:560- 571.
Llinas, Rodolfo R., and R. Hess (1976). Tetrodotoxin -resistant dendritic spikes in avian
Purkinje cell . National Academy of Sciencesof the United States of America . Proceedin(~s.
Biological Sciences. 73:2520- 2523 .
Llinas , Rodolfo R., and D . E. Hillman (1969). Physiological and morphological organization of the cerebellar circuits in \'arious vertebrates. In Neurobiologyof cerebellar
evolutionand development
, ed. R. Llinas , 43- 73. Chicago: American Medical Association
in vitro . Nature
Llinas , Rodolfo R., and M . Sugimori (1979). Calcium conductances in Purkinje cell
dendrites : Their role in development and integration . In Progressin Brain Research
.
Vol . 51: Developmentand chemicalspecificity of neurons, ed. M . Cuenod, G. \Atr
.
Kreutzberg , and F. E. Bloom, 323- 334. Amsterdam : Elsevier.
Llinas, Rodolfo R., and M . Sugimori (1980). Electrophysiological properties of in vitro
510 Bibliography
Purkinje cell somata in mammalian cerebellar slices. Journal of PhysiologyLondon
305 :171 - 195 .
University
Press .
Lynch , Gary , James L . McGaugh , and Norman M . Weinberger, eds. (1984). Neurobiology of learning and memory. New York : Guilford .
McCarthy , J., and P. Hayes (1969). Some philosophical problems from the standpoint
of Artificial Intelligence . In Machine intelligence, vol . 4, ed. B. Meltzer and D.
Michie , 463- 502 . New York : Harper and Row .
(1981) . An interactive
activation
model
Press
McClelland , JamesL ., David E. Rumelhart , and Geoffrey E. Hinton (1986). The appeal
of parallel distributed processing. In D . E. Rumelhart and J. L . McClelland , eds.
(1986). Cambridge , Mass.: MIT Press.
McCloskey , Michael (1983). Intuitive physics . ScientificAmerican 248/4:122- 130.
McCloskey , Michael , Alfonso Caramazza, and Annamaria Basili (forthcoming ). Dissociations
of number
system
neurophysiological
perspective
, ed. G. Eeloshe and X. Seron. Hillsdale , N .J.: Erlbaum .
McCloskey , Michael , Alfonso Caramazza, and Bert Green (1980). Curvilinear motion in
the absence of external forces: Naive beliefs about the motion of objects. Science
210 / 4474
: 1139 - 1141
McCollum , G., A . Pellionisz , and R. Llinas (1983). Tensorial approach to color vision .
Journal of TheoreticalNeurobiology2:23- 28.
McCormick , David (1984). Cerebellum : Essential involvement in a simple learned response. Doctoral dissertation , Stanford University .
McDowell , Fletcher (1979). Cerebrovascular diseases: Section fourteen . In Textbookof
medicine, ed. Paul B. Beeson, Walsh McDermott , and James B. Wyngaarden , 777801. Philadelphia : Saunders.
McEwen , Bruce S. (1976). Interaction between hormones and nervous tissues. Scientific
American
235 / 1 :48 - 58 .
McEwen , Bruce S., Carl J. Denef, John L. Gerlach, and Linda Plapinger (1974). Chemical studies of the brain as a steroid
Worden , eds . (1974), 599- 620 .
and F . G .
McGinn , Colin (1982). The characterof mind. Oxford : Oxford University Press.
MacKay- , D . M . (1978) . Selves and brain . Neuroscience 3:599- 606 .
511
Bibliography
MacVicar
, B ., and
slices
F . E . Dudek
of rat hippocampus
Magendie
. Brain
between
, F . J. ( 1822a ) . Experiences
des racines
, F . J. ( 1822b ) . Experiences
de la moelle
lated
in E . Clarke
Magistretti
, Pierre
System
Makowski
epiniere
C . D . O ' Malley
J. , John
H . Morrison
, Lee , D . L . D . Caspar
, and
and
des nerfs
translated
des racines
Floyd
. II : Analysis
des nerfs
of X - ray
, and
1/ 2 (special
D . A . Goodenough
diffraction
( 1976 ) . Cooperative
, Takayoshi
, and
. Tokyo
Asao
/ New
Hirano
, J. C . , M . E . Phelps
of human
of nonreductive
computation
to brain
of stereo
development
by electrical
stimulation
Bulletin
13 :94 - 100 .
: Principles
, R . E . Carson
cerebral
metabolism
, and
Merzenich
, M . M . , and J. F . Brugge
Metropolis
plane
, M . M . , and
perceptual
systems
ogy 9 :1- 42 .
disparity
J. H . Kaas
calculations
. Science
feedback
functions
in
-
by
Richard
, M . Rosenbluth
for
fast
computing
in Neuro -
( 1982 ) . Tomographic
. Neurology
Jed Wyatt
( 1977 ) . Human
32 :921 sleep
monkey
( 1980 ) . Principles
. Progress
to morphine
stimulation
? (Commentary
( 1973 ) . Representation
of the macaque
in mammals
, N ., A . Rosenbluth
tion of state
6 : 1087 .
materialisln
. Discussions
D . E . Kuhl
: Auditory
temporal
Biology
: Comparison
and applications
the superior
issue ) .
Mendelson
, Wallace
B . , J . Christian
Gillin , and
and its disorders . New York : Plenum .
Merzenich
trans -
Mazziotta
, J. C . ( 1985 ) . PET scanning
sciences 2 :9 - 47 .
mapping
937 .
and
specialization
: What , how , and why . (Commen
1981 .) Behavioral and Brain Sciences 4 / 1 :72 - 73 .
as guides
Mazziotta
qui naissent
of Cell
Mateer , Catherine
A . ( 1983 ) . Localization
of language
and visuospatial
electrical
stimulation
. In A . Kertesz , ed . ( 1983b ) , 153 - 183 .
tomography
and
( 1977 ) . Gap
data . Journal
Marsden , C . D . , J. C . Rothwell
, and B . L . Day ( 1984 ) . The use of peripheral
the control
of movement
. Trends in Neurosciences 7 :253 - 257 .
Matsui
rachidiens
in E . Clarke
( 1969 ) . A theory
of cerebellar
cortex . Journal of Physiology
( 1982 ) . Vision . San Francisco : W . H . Freeman .
in
in Neurosciences
, W . C . Phillips
Margolis , Joseph
( 1978 ) . Persons and
Dordrecht
, Holland
: D . Reidel .
Marr , David
Marr , David
cells
junction
structures
74 :629 - 645 .
pyramidal
and
Development
CA3
of the cochlear
. Brain
of
, A . Teller , and
machines
partition
on
organization
in Psychobiology
on Dennett
of
and Physiological
E . Teller
. Journal
sensory
Psychol -
( 1953 ) . Eaua -
of Chemical
Physics
512
Bibliography
Milner , Brenda (1966). Amnesia following operation on the temporal lobes. In Amnesia,
ed. C. W . M . Whitty and O . Zangwill , 109- 133. London : Butterworth .
Milner , B., S. Corkin , and H .-L . Teuber (1968). Further analysis of the hippocampal
amnesic syndrome : 14-year follow -up study of H . M . Neuropsychologia
6:215- 234.
Milner , B., and M . Petrides (1984) . Behavioral
Trends
in Neurosciences
7/ 11 :403 - 407 .
Milner , B., and T. Rasmussen (1966). Evidence for bilateral speech representation in
some non-right handers. Transactionsof the American NeurobiologicalAssociation
91 :306 - 308 .
Mishkin , Mortimer , Leslie G. Ungerleider , and Kathleen A . Macko (1983). Object vi sion and spatial vision : Two cortical pathways . Trendsin Neurosciences
6:414- 417.
Mohr , J. P. (1973). Rapid ameliorization of motor aphasia. Archivesof Neurology28:7782 .
of
the
extrastriate
cortex
. Science
Morrell , Joan I ., and Donald W . Pfaff (1981). Autoradiographic technique for steroid
hormone localization : Application to the vertebrate brain . In Neuroendocrinologyof
reproduction , ed . Norman
system . In F. O . Schmitt
Bibliography
513
l\' aeser, A . M ., M . Alexander , N . Helm -Estabrooks, et al. (1982). Aphasia with predominantl.y subcortical lesion sites. Archives of Neurology 39:2- 14.
Nagel, Ernest (1961). The structlJreof science
. New York : Harcourt , Brace, and World .
Nagel, Thomas (1974). What is it like to be a bat? PhilosophicalRevieu' 83:435- 450.
Neher , E., and C. F. Stevens (1979). Voltage-driven conformational changes in intrinsic
membrane proteins . In F. O . Schmitt and F. G. Worden , eds. (1979), 623- 629.
Neisser, Ulric (1966). Cognitive psychology. New York : Appleton -Century -Crofts .
Neuburger , Max (1897). Die historischeEnhvicklun:l, der cxperil1lcntellcnGehirlz- LindRi/('ken1narksphysiologie
vor Floure11s
. Stuttgart : Ferdinand Enke Verlag. (English translation by Edwin Clarke (1981). Thehistoricaldevelopmentof experimentalbrain alzdspinal
cord physiologybeforeFlourens. Baltimore : The Johns Hopkins University Press.)
Neurath , Otto (1932). Protokollsatze . Erkenntnis 3:204- 214. (Translated by G. Schlick
and reprinted as " Protocol sentences" in A . J. Ayer , ed. (1959). Logicalpositiz'isln.
New York : Macmillan .)
Nicholls , J. G., and D . A . Baylor (1968). Specific modalities and receptive fields of
sensory neurons in the CNS of the leech. Journal of Neurophysiology31:740- 756.
Nicoll , Roger A . (1982). Neurotransmitters can say more than just " yes" or " no ."
Trends in Neurosciences
5:369- 374.
Nisbett , Richard, and Lee Ross (1980). Human inference
: Strategiesand shortcomingsof
socialjudgment. Englewood Cliffs , N .J.: Prentice-Hall .
Norman , D . A . (1969). Memory and attention: An introduction to hull1aninforlnation process
ing. New York : Wiley- .
Norman , D . A . (1973). Memory , knowledge , and the answering of questions. In The
Loyolasy,nposiumon cognitivepsychology
, ed. R. Salsa, 135- 165. Washington , D .C.:
Winston .
Norman , D . A ., and D . E. Rumelhart (1970). A system for perception and memory . In
Models of human memory, ed. D. A . Norman . New York : Academic Press.
Northcutt , R. Glenn (1977). Nervous system (vertebrate). In McGrazv-Hill encyclopedia
of
scienceand technology, vol . 9, 90- 96. New York : McGraw -Hill .
Nottebohm , F. (1981). Laterality , seasonsand space governing the learning of a motor
skill . Trends in Neurosciences
4/5:104- 106.
Ojemann , George A . (1983). Brain organization for language from the p,~rspective of
electrical stimulation mapping . Behavioraland Brain Sciences6/2:189- 230.
O'Keefe, J., and L. Nadel (1978). Thehippocampusasa cognitivemap. Oxford : Clarendon
Press.
alton , D ., J. T. Becker, and G. E. Handelmann (1979). Hippocampus , space, and
memory . Behavioraland Brain Sciences2:313- 365.
Oppenheim , P., and H . Putnam (1958). Unity of science as a working hypothesis . In
Minnesota studies in the philosophyof science
, vol . 2, ed. H . Feigl et al., 3- 36. Min neapolis: University of Minnesota Press.
Palm, Gunther (1982). Neural assemblies
: An alternative approachto artzfi'cial intelli (,ence.
Berlin : Springer-Verlag .
Patterson, Paul H ., and Dale Purves, eds. (1982). Readingsin develop,nental neurobiology
.
Cold Spring Harbor , N .Y.: Cold Spring Harbor Laboratory .
Paupardin -Tritsch , D ., and H . M . Gerschenfeld (1975). 2-hydroxytryptamine as a syn.
aptic transmitter of excitation and inhibition . In Golgi centennialsymposiumproceed
ings, ed. M . Santini , 487- 501. New York : Raven.
Peirce, Charles Sanders (1931- 1935). ThecollectedpapersafCharlesSandersPeirce, vols. 16, ed. C. Hartshorne and P. Weiss. Cambridge, Mass.: Harvard University Press.
Pellionisz , Andras (1983a). Brain theory : Connecting neurobiology to robotics. Tensor
514 Bibliography
analysis: Coordinates to describe, understand and engineer functional geometries
of intelligent organisms . Journal of TheoreticalNeurobiology2:185- 211.
Pellionisz , Andras (1983b). Sensorimotor transformations of natural coordinates via
neuronal networks : conceptual and forma ] unification of cerebellar and tectaI models. In Proceedingsof workshopon visuomotorcoordinationin frog and toad. (Organized
by R. Lara and M . Arbib .) Computer and Information Science: University of Massachusetts Technical Report 83- 19.
Pellionisz, Andras (1984). Coordination : A vector-matrix description of transformations
of overcomplete CNS coordinates and a tensorial solution using the MoorePenrose generalized inverse . Journal of TheoreticalBiology 110:353- 375.
Pellionisz, Andras (1985). Tensorial aspects of the multidimensional approach to the
vestibulo -oculomotor reflex and gaze. In Adaptive mechanismsin gazecontrol, ed. A .
Berthoz and G. Melvill Jones, 231- 296. Amsterdam : Elsevier.
Pellionisz , A ., and R. Llinas (1979). Brain modeling by tensor network theory and
computer simulation . The cerebellum : Distributed processor for predictive coordination . Neuroscience4:323- 348.
Pellionisz, A ., and R. Llinas (1982). Space-time representation in the brain . The cerebellum as a predictive space-time metric tensor. Neuroscience7:2249- 2970.
Pellionisz, A ., and R. Llinas (1985a). Cerebellar function and the adaptive feature of the
central nervous system. In Adaptive mechanismsin gazecontrol, ed. A . Berthoz and
G. Melvill Jones, 223- 232. Amsterdam : Elsevier.
Pellionisz, A ., and R. Llinas (1985b). Tensor network theory of the metaorganization of
functional geometries in the central nervous system. Neuroscience16:245- 273.
Pellionisz , A ., R. Llinas , and D . H . Perkel (1977). A computer model of the cerebellar
cortex of the frog . Neuroscience2:19- 36.
Penfield, Wilder , and Herbert Jasper (1954). Epilepsyand the functional anatomy of the
human brain. Boston: Little , Brown and Co.
Penfield, Wilder , and Lamar Roberts (1959). Speech
and brain mechanism
. Princeton, N .J.:
Princeton University Press.
Perecman, Ellen, ed. (1983). Cognitive processingin the right hemisphere
. New York : Academic Press.
Perenin, M . T., and M . Jeannerod (1978). Visual function within the hemianopic field
following early cerebral hemidecortication in man. I . Spatial localisation . Neuropsychologia16:1- 13.
Perenin, M . T., J. Ruel, and H . Hecaen (1980). Residual visual capacities in a case of
cortical blindness . Cortex 16:605- 612.
Perry, Elaine K ., and Robert H . Perry (1982). The cholinergic system in Alzheimer ' s
disease. Trends in Neurosciences
5:261- 262.
Peterson, B., J. Baker, C. Wickland , and A . Pellionisz (1985). Relation between pulling
directions of neck muscles and their activation by the vestibulocollic reflex: Tests of
a tensorial model . Societyfor NeuroscienceAbstracts 11:83.
Pfaff, Donald W . (1980). Estrogensand brain function. New York : Springer-Verlag.
Pfaff, Donald , W ., and Bruce S. McEwen (1983). Actions of estrogens and progestins
on nerve cells. Science219:808- 814.
Pflug , B., and R. Tolle (1971). Therapy of endogenous depression by sleep deprivation :
Practical and theoretical consequences. Nervenarzt 42:117- 124.
Phelps, Michael E., David E. Kuhl , and John C. Mazziotta (1981). Metabolic mapping of
the brain 's response to visual stimulation : Studies in humans . Science211:14451448.
Phelps, Michael E., and John C. Mazziotta (1985). Positron emission tomography :
Human brain function and biochemistry . Science228:799- 809.
515
Bibliography
Piaget, Jean (1971). Insights and illusions of philosophy. Translation by Wolfe Mays. New
York : World Publishing Co.
Picton, T. W ., and S. A . Hillyard , H . I . Krausz, and R. Galambos (1974). Human
auditory evoked potentials . I . Evaluation of components . Electroencephalography
and
Clinical Neurophysiology36:179- 190.
Pitts , W . H ., and W . S. McCulloch
of
Foundation
. In G .
Poulos, Constantine X., and D . Adrian Wilkinson (1984). A process theory of remembering : Its application to Korsakoff amnesia and a critique of context and episodicsemantic theories . In L . R . Squire and N . Butters , eds . (1984), 67- 82.
Puccetti, Roland (1981). The casefor mental duality : Evidence from split -brain data and
other
considerations
. Behavioral
and Brain
Sciences 4 / 1 :93 - 99 .
Pylyshyn , Zenon (1980). Computation and cognition : Issues in the foundation of cognitive
science . Behavioral
and Brain
516 Bibliography
Quine , W . V . O . (1969). Epistemology naturalized . In Ontological relatir1ityand other
essays, 69- 90 . New York : Columbia
University
Press .
Quinn , William G., and Ralph J. Greenspan (1984). Learning and courtship in
Drosophila: T~!o stories with mutants . Annual Revieu1of Neuroscience7:67- 93.
Raichle, Marcus E., and Michel M . Ter-Pogossian (1984). Letter in reply to Fox: PET
scan controversy
Raisman, G., and P. M . Field (1973). Sexual dimorphism in the neuropil of the preoptic
area of the rat and its dependence on neonatal androgen . Brain Research54:1- 29.
Rakic, Pasko (1975). Cell migration and neuronal ectopias in the brain . Birth Defects:
Original Article Series11:95- 129.
Rakic, Pasko (1981). Developmental events leading to laminar and areal organization of
the
neocortex
. In
F . O . Schmitt
, F . G . Worden
, G . Adelman
, and
S . G . Dennis
eds . (1981), 7- 28 .
Ramon y Cajal, Santiago (1888). Estructura del cerebelo. Gac. med. Catalana11:449- 457.
(Excerpted and translated in E. Clarke and C. D . O'Malley (1968), 111- 113.)
Ramon y Cajal, Santiago (1908). Structure et connexions des neurones. LesPri.t Nobelen
1906, 1- 25 . Stockholm : P. A . Norstedt
Clarke and C . D . O ' Malley
in E.
Ranson, Stephen W ., and Sam L . Clark (1959). Theanatomyof the nervoussystem. Philadelphia : Saunders.
Rasmussen, T., and B. Milner (1977). The role of early left-brain injury in determining
lateralization of cerebral speech functions . Annals of theNeu' York Academyof Science
299 :355 - 369 .
Richards, Whitman (1983). Is hindsight better than blindsight ? (Commentary on Campion , Latto , and Smith 1983 .) Behavioral and Brain Sciences 6:461 .
of selective attention
in animals . In Advances in
vertebrateneuroethology
, ed. J.-P. Ewart , R. R. Capranica, and D . J. Ingle , 261- 297.
New
York : Plenum
development
of
Rorty , Richard (1965). Mind -body identity , privacy , and categories. Revie{{, of Metaphysics19:24- 54. (Reprinted in D . M . Rosenthal, ed. (1971), 174- 199.)
Rorty , Richard (1970). In defense of eliminative materialism . Reviezvof Metaphysics
24:112- 121. (Reprinted in D . M . Rosenthal, ed. (1971), 223- 231.)
Rosch, E. (1973). On the internal structure of perceptual and semantic categories. In
Bibliography
517
Cognitive developmentand the acquisitionof language, ed. T. Moore, 111- 144. New
York : Academic Press.
Rosch, E., and B. B. Lloyd , eds. (1978). Cognition and categorization
. Hillsdale , N .J.:
Erlbaum .
Rosch, E., and C. Mervis (1975). Family resemblances: Studies in the internal structure
of categories. Cognitive Psychology7:573- 605.
Rose, Clifford F., and W . F. Bynum (1982). Historical aspectsof the neurosciences
. New
York : Raven.
Rosenberg, Jay (1974). Linguistic Representation
. Dordrecht , Holland : D . Reidel.
Rosenberg, Roger N . (1980). Genetic variation and neurological disease. Trends in
Neurosciences3/6:144- 148.
Rosenblatt, Frank (1962). Principlesof neurodynamics
. Washington , D .C.: Spartan Books.
Rosenthal, D . M ., ed. (1971). Materialism and the mind-bodyproblem. Englewood-Cliffs ,
N .J.: Prentice-Hall .
Rubens, Alan B., and D . Frank Benson (1971). Associative visual agnosia. Archives of
neurology24:305- 316.
Rumelhart , David E., and Donald A . Norman (1982). Simulating a skilled typist : A
study of skilled cognitive -motor performance . Cognitive Science6:1- 36.
Rumelhart , David E., and James L. McClelland (1986). Parallel distributed processing
.Explorationsin the microstructureof cognition. Vol . 1: Foundations. Cambridge, Mass.:
MIT Press.
Rumelhart , D . E., P. Smolensky, J. L. McClelland , and G. E. Hinton (1986). Schemata
and sequential thought processes in POP models. In J. L. McClelland and D . E.
Rumelhart , eds. (1986).
Rumelhart , David E., and David Zipser (1985). Feature discovery by competitive learning . Cognitive Science9:75- 112.
Russell, Bertrand (1914). On the nature of acquaintance. In Logic and knozvledge
(1956),
ed. R. Marsh , 127- 174. London : Allen and Unwin .
Russell, Bertrand (1924). Logical atomism . In Logicand knozL11edge
(1956), ed. R. Marsh ,
321- 343. London : Allen and Unwin .
Russell, Bertrand, and A . N . Whitehead (1910- 1913). Principia mathematica
. 3 vols .
Cambridge : Cambridge University Press.
Sacks, Oliver (1985). The ty..rins. New York Reviezvof Books32:16- 20.
Salmon, Wesley (1971). Statisticalexplanationand statistical relevance
. Pittsburgh : Univer sity of Pittsburgh Press.
Saunders, J. B. de C. M ., and C. D. O'Malley (1950). The illustrations from the u'orks of
Andreas Vesalius. Cleveland : The World Publishing Co.
Schank, Roger, and Robert Abelson (1977). Scripts, plans, goalsand understanding. Hills dale, N .J.: Erlbaum .
Schiller, P. H . (1984). The superior colliculus and visual function . In Handbooko/ physiology. SectionI: Thenervoussystem, ed. I. Darian-Srnith, vol. 3, 457- 504. Bethesda, Md .:
American Physiological Society.
Schiller, P., and M . Stryker (1972). Single-unit recording and stimulation in superior
colliculus of the alert rhesus monkey . Journal of Neurophysiology35:915- 924.
Schlick, Moritz (1936). Meaning and verification . PhilosophicalReviezv46:339- 369.
Schmitt, F. 0 ., and F. G. Worden , eds. (1974). The neurosciences
: Third study program.
Cambridge, Mass.: MIT Press.
Schmitt, F. 0 ., and F. G. Worden , eds. (1979). The neurosciences
: Fourth study program.
Cambridge, Mass.: MIT Press.
Schmitt , F. 0 ., F. G. Worden , G. Adelman , and S. G. Dennis , eds. (1981). Theorganization of the cerebralcortex. Cambridge , Mass.: MIT Press.
518
Bibliography
Bibliography
519
, G . Maxwell
, and
I . Savodnik
, 163 - 177 .
SperryI R. W . (1980). Mind -brain interaction : Mentalism , yes; dualism , no. Neuroscience
5 :195 - 206 .
Sperry, R. W ., and Michael Gazzaniga (1967). Language following surgical disconnection of the hemispheres . In Brain mechanismsunderlying speechand language, ed. C.
Millikan and F. Darley , 108- 115. New York : Grune & Stratton .
Spitzer , Nicholas
4 / 5 : 115 - 118 .
Springer, Sally P. (1977). Tachistoscopic and dichotic -listening investigations of laterality in normal human subjects. In S. Harnad et al., eds. (1977), 325- 336.
Springer, Sally P., and Georg Deutsch (1981). Left brain, right brain. San Francisco:
W . H . Freeman
Squire, Larry R., and Neal J. Cohen (1984). Human memory and amnesia. In G. Lynch ,
J. L . McGaugh , and N . M . Weinberger, eds. (1984), 3- 64.
Stent , Gunther
circuits
generating
Press .
" hypercolumn " size in areas 3b and 1 of somatosensory cortex in owl monkeys .
Journal of Neurophysiolofl ..1/ 44 :295- 311 .
Sutton , 5., P. Teuting , J. Zubin , and E. R. John (1967). Information delivery and the
sensory evoked potential . Science 155:1436- 1439.
520 Bibliography
Boerhaave. Amsterdam : Severinus. Translated into English by T. Flloyd (1758).
London
of seman -
Taylor , Charles (1971). Interpretation and the sciences of man . Revieu1of Metaphysics
25:1- 32, 35- 45. (Reprinted in Charles Taylor (1985). Philosophyand the human sciences
: Philosophicalpapers, vol . 2. Cambridge : Cambridge University Press.)
Taylor , P. A ., and T. W . Stone (1981). Neurotransmodulatory control of cerebral cortical neuron activity . In F . O . Schmitt , F . G . Worden , G . Adelman , and S. G .
Dennis , eds . (1981), 347- 357 .
Ter-Pogossian, Michel M ., Marcus E. Raichle, and Burton E. Sobel (1980). Positronemission tomography . ScientificAmerican 243:171- 181.
Thompson , R. F. (1967). Foundationsof physiologicalpsychology
. New York : Harper and
Row
Row .
(1982) . Illusory
conjunctions
in the perceptions
of
Truman , JamesW . (1984). Cell death in invertebrate nervous systems. Annual Reviewof
Physiology7, 171- 188.
Tulving , Endel (1972). Episodic and semantic memory . In Organizationof memory, ed. E.
Tulving and W . Donaldson . New York : Academic Press.
Tulving , Endel (1983). Elementsof episodicmemory. Oxford : Clarendon Press. (Precis
with commentaries and replies in Behavioraland Brain Sciences(1984), 7:223- 268.)
Turing , A . M . (1937a). On computable numbers, with an application to the Entscheidungsproblem . Proceedingsof the LondonMathematicalSociety42:230- 265.
Turing , A . M . (1937b). Computability and lambda-definability . Journalof SymbolicLogic
2 : 153 - 163 .
Ullman , Shimon (1979). Interpretation of visual motion. Cambridge, Mass.: MIT Press.
Bibliography
521
Uttal , William R. (1973). Thepsychobiologyof sensorycoding. New York : Harper and Row .
Valenstein, Elliot S. (1973). Brain control. New York : Wiley .
Van Essen , David C ., and John H . R . Maunsell
functional
streams
in the visual
(1983) . Hierarchical
cortex . Trends
in Neurosciences
organization
and
6 :370 - 375 .
Van Fraassen, Bas C. (1970). On the extension of Beth's semantics of physical theories.
Philosophy of Science 37:325- 338 .
, Holland
: Reidel
Press .
Vesalius, Andreas (1543). The Epitomeof the De Humani Corporis. Brussels. Reproduced
in J. B . de C . M . Saunders and C . D . O ' Malley (1950).
Vignolo , Luigi A . (1983) . Modality
ed . (1983b ), 357- 369 .
language . In A . Kertesz ,
in human
3 :5 .
Volpe , Bruce T., John J. Sidtis, and Michael S. Gazzaniga (1981). Can left -handed
writing posture predict cerebral language laterality ? Archives of Neurology 38:637638 .
neuron
responses
Von der Malsburg , Christoph (1973). Self-organizing of orientation sensitive cells in the
striate cortex. Kybernetik14:85- 100.
Von der Malsburg , Christoph (1981). Internal report 81- 2. Gottingen , F. R. G.: Department of Neurobiology , Max-Planck-Institute for Biophysical Chemistry .
Von der Malsburg , Christoph , and David Willshaw (1981). Cooperativity and brain
organization
Von Neumann , John (1958). The computer and the brain . New Haven , Conn . : Yale Uni versity Press .
Waber , D . (1976) . Sex differences
in cognition : A function
of maturation
rate ? Science
Wada, J. (1949). A new method for the determination of the side of cerebral speech
522
Bibliography
dominance . A preliminary report on the intracarotid injection of sodium amy tal in
man. Medical Biology 14:221- 222.
(1960) . Intracarotid
injection
of sodium
lateralization of cerebral speech dominance . Experimental and clinical observations . Journal of Neurosurgery17:266- 282.
Waldeyer -Hartz , H . W . G. von (1891). Uber einige neuere Forschungen im Gebiete der
Anatomie des Centralnervensystems . Dt . med. Wschr. 17:1213- 1218, 1244- 1246,
1267- 1269, 1287- 1289, 1331- 1332, 1352- 1356. (Excerpted and translated in E.
Clarke and C . D . O ' Malley
syn -
C . Thomas
and
Oxford
: Basil
Blackwell
neurons
York : Plenum
ity in the hemianopic field following a restricted occipital ablation . Brain 97:709728 .
and
senile
: 1237 - 1239
dementia
: Loss
of neurons
in the basal
forebrain
. Science 215 /
Willshaw , David (1981). Holography , associative memory , and inductive generalization . In G . E . Hinton
Wimsatt , W . (1976b). Reduction , levels of organization , and the mind -body problem .
In
Consciousness
267 . New
and
the
York : Plenum
brain
, ed . G . Globus
, G . Maxwell
, and
I . Savodnik
, 199 -
Witelson , S. F. (1976). Sex and the single hemisphere : Specialization of the right hemisphere for spatial processing. Science193:425- 427.
Wittgenstein , Ludwig (1922). Tractatuslogico-philosophicus
. Modern edition with transla-
Bibliography
523
Index
Amphetamine psychosis, 83
Amplification , neuronal , 65
Amygdala , 102, 106
and
localization
and
memory
and
testosterone
, 171 - 172
' , 369
, 90
propositions
Anarithmia
, 244
, 233
Anastomosis
, 27
320
Anions
, 49
Anterior
commissure
, 174 , 176
Functionalism
functional
and
intertheoretic
and operations
theories
reduction
, 297
on representations
297 - 299
in physicalism , 347
Anton ' s syndrome . See Blindness
Anubis
baboons
denial
, 96
nicke' s aphasia
526 Index
Aplysia Califarnica (sea hare ), 67, 70- 71,
72 , 79 , 369 , 372 , 488n9
. 10
Bacon
, SirFrancis
, 358
Ballard
, D. H., 460, 461
Barbary macaque, 96
Arbib
, Michael
Aristarchus
, 488n10 .3
, 240 - 241
Aristotle
Arithmetic
, Roberts
, 166 - 167
Behavior
and
, H . , 133
Association
Bartholow
conditions
as level of organization
Behaviorism
, 359
, 278
As trocytes , 42
Ataxia
, 413
and
co -evolution
and
ERPs
, 285
, 210 , 216
and intentionality
, 343- 344
, 20 - 21
Bell - Magendie
Law , 20
receptive
structure
Berlucchi
Bianchi
, 232
Autonomy
rationale
Awareness
377 - 380
Blindness
consciousness
Blindness
dualism
, 486
deficits
hillock
denial
Blood -brain
, 172
, 54 , 55
Axons
discovery of, 27
function
, 309
of mental
Axon
, 463 , 472
Bisiach , Edoardo
as construct
, 163
and
, Leonardo
assessment
, G ., 115
of , 38
barrier
, 86
Botulinus
nons piking , 59
processes
Bou
of , 28
tons
toxin , 67 - 68
. See End
bulbs
~-endorphin , 79
Baboons, 96
androgenization
blood
vessels
of , 89- 91, 92
in , 155 , 156
Index
cytoarchitectural map of, 102, 105
divisions of, 100, 101 (seealso specifi
'c
structures )
functional
architecture
disorders
eases
, as neurochemical
cortex , 414
Cerebellum
and
Cerebellar
cerebral
evolution
of , 104
, 85 - 86
cortex , 414
of , 413 - 414
function
dis -
of , 413
micro -organization
Brentano
neuronal
, Franz
, 336
of , 412- 413
network
of , 431
Cerebral
Broca
' s area
259 - 260
, Korbinian
Brodmann
, 102
cortex ; Pri -
Chemical
synapses
Chick , 140 , 369
, T . H . , 209
Burklund
Chimeric
, C . W . , 188
California
" Caloric
C - fibers , 41
Brodmann
stimuli
, 62 , 65
Chimpanzee , 168
Chloride ions (CI - ), 48, 52
Chlorpromazine , 82, 83
Cholecystokinin , 78
Cholinergic neurons, 78
Christensen
, B ., 100
Churchland
, Paul
, neuronal
, 69 - 76 , 414 , 415
Canary , 94
for cerebellar
Carmine
as level of organization
stain , 27
, Rene
cortex , 414
, 359
as theory , 408
Classical conditioning , 71- 72
Classical empiricism , 244- 247
Classical
Cat , 176
mechanics
302
527
, 48 , 49
448
Causal relations
Cell assemblies
Cell
, 48 - 50
Cocaine
, 82
Cochlea
, 125 , 126
membrane
excitability of, 57
potential
difference
across , 51
Cochlear
nuclei
, 125
Cellular
Co - evol u tion
ts of , 363 - 364
and
electrical
synapses
, 62
528
Index
Co -evolution
(continued )
and
reduction
376
364
two - dimensional
order
in , 133
and
, 181
Commonsense
chology
consciousness
, 460
as , 264 - 265
substrates
, 360
structures
, 360 - 361
organization
, 352- 353,
354 - 355
skills
cognition
as, 9, 350
as coordinate
transformation
488n10 . 1, 488nl0
, 426 , 440 ,
.7
and
neuroscience
, 355
functionalism
and information
storage , 458 - 460
and tn -level organization , 359, 360
and word -nonword recognition , 459460
Computer modeling
of frog nervous system, 416, 439, 440,
441
and
psy -
350
and
skills , 449
Cohen
, NealJ., 371
Cole, J., 171
Coloragnosia
, 172
for gaze
control
, 437 - 439
" commonsense
Index
Conceptanalysis, method of, 271- 274
Concepts
, mental. SeealsoFolkpsychology
as re,risable, 275
and theory of mind, 301
Conceptual-role theory. SeeNetwork
theory of meaning
Conditionedstimulus, 71
Conduction, 41- 43, 61
Conductionaphasia, 161
Cones, and "lightness" values, 454
Confabulation
in anterogradeamnesia, 371
in blindnessdenial, 228
Confirmationtheory, 252, 265, 271
Connectionism
, 412, 458- 474
and Crick- Mitchison hypothesis,
484n5.1
manifestoof, 461
and tensornetwork theory, 448
Connectivity, synaptic, and tensor network theory, 428, 430, 432
Consciousness
, 152. SeealsoAwareness
as accidentalkind, 321
and antireductionistarguments, 319
function of, 208- 209
and left hemisphere, 180- 181
and memory, 372
as monitor, 321
unity of, 173, 178- 182, 321- 322
Constantine- Paton, Martha, 140- 141
Constraints, on brain theory, 409- 410
Contralateralinnervation, 166, 186, 187
Contrastenhancement
, 73, 75, 76
Control, unity of, 173, 178- 182, 371
Convergence
, and neuronsin an array,
416
Cooperativesearch. Seea/soHintonSejnowskimodel
for "best hypothesis," 463
constraintnetwork for, 465
and energyminima, 467
and figure/ground ambiguity, 472
relaxationstatesof, 466
Coordinatespace. SeePhasespace
Coordinatetransformation, 412, 418,
423, 425, 430, 461, 464- 467
computationas, 426, 440, 488n10
.1
by contiguoustopographicmaps, 447
by
428, 429
- matrix multiplication,
and phasespacedeformation, 424- 425,
426
529
, Suzanne
, 370
information
transfer
across , 197
horizontal
lesions
in , 95 - 96
stratification
of (see Laminae )
of , 228
Cranioscopy . SeePhrenology
Creativity , 183
Crick , F . H . C . , 125 , 208 , 285 , 315
induction
, 250
Axial
Tomography
Cummins
, Robert
, 294
, and
' s law
Darwin
chemical
synapse
, 67
, 78
, Charles
Deafferentation
, 253 , 263
, 431 - 432
ti ve memory
, 484n3 . 1
Deductive
logic , 261
, Otto
F . K . , 27
, W . C . , 207 - 208
Democritus
, 240
530
Index
Dendrites
Dopaminergic
neurons
Dreaming
206
connections
Drosophila
369
of , 263
discovery
function
of , 27
Drugs
of , 38
processes
spiking
of , 28
in
and
split
neurotransmitters
, 64
substance
dualism
, 318 - 319
, 463 , 464
Pierre
in
180
Emil
260
318
meaning
261
320
337
182
24
262
413
413
Early
selection
216
Eastern
series
Eccles
EEG
theories
and
John
and
. ,
neural
291
204
and
for
sleep
mapping ,
129 - 130
Efferent
neurons
development
of
,
136 - 137
Albert
. 2
) ,
207
-
204
40
138
210
209
117
139
253
260
262
55
263
264
286
Electrical
gradient
Electrical
potential
49
,
Action
potentials
tentials
Electrical
waveforms
209
254
,
54
defined
Synaptic
Event
50
See
a / so
related
po
potentials
stimulation
studies
165
168
171
synapses
Electricity
"'
also
and
62
nerve
Electrical
function
24
See
210
potentials
( EEG
) ,
204
204
radiation
oretic
reduction
,
force
Electron
279
,
microscopy
Electroreceptors
437
,
209
stages
Electromotive
data
205
206
Electromagnetic
Distortion , in somatosensory
486n8
Electroencephalograph
190
186
,
basis
Electrical
Dimorphism , 88
Disconnection effect , in split -brain stud -
ERP
( quoted
319
potentials
dreaming
Eigenpsychische
, 164
176
action
Einstein
Diaschisis
and
( electroencephalograph
DeVoogd, T. J., 94
, E . M . , 365
217
Eigenposition
Dominance
408
of
,
,
Dysmetria
Dewan
Donchin
Reymond
Neuroleptics
studies
characteristic
Deterministic
strategy , and Hinton Sejnowski model , 468
Deutsch , Georg , 197
model
See
sleep
83
REM
substances
- brain
Bois
Duhem
also
316
theory
Dysarthria
du
78
See
metaphor
Dennett
of
network
278
computer
and
and
3 ,
interaction
208
antipsychotic
Dualism
of , 59, 60
Elimination
See
and
31
interthe
280
50
,
,
30
33
63
47
Ontological
simplification
Eliminative
materialism
Emergence
323
,
-
327
and
377
mental
395
states
399
4 ,
317
Index
Emergent property , as " network prop erty ," 324- 325
Empirical propositions , 244
Em piricism . See Classical empiricism ;
Logical empiricism
Empiricist
En <; , Berent
, 357
Endorphins , 69, 78
Energy, and Hinton - Sejnowski model ,
467 - 470 , 489n10
Enkephalins , 69, 78
Ependymal cells, 42
Epilepsy
and commissurotomy , 173, 176
and
dominance
and
EEG
tests , 194
traces , 204
, 20
substrate
of , 213 - 216
, 78 , 91 , 92
Estrogen , 89, 91
Ethology , and unified
Evolution
substance
dualism
, 320
postsynaptic
potentials
(EPSPs). SeeDepolarization
Exogenous components , in ERPs, 210,
214
exper -
ience
Experimental
531
Explanation
deductive- nomologicalmodel of, 294,
303- 304
and intertheoreticreduction, 294, 296
and logical empiricism, 255- 256, 258,
265
in neurophysiology, 296, 297, 378
in psychology, 303- 304, 378
Extracellularfluid, 48
Extraocularmuscles, 434
Extrapolator-deducermechanism
, 391
Eyeballrotation, 435- 436
Facial agnosia, 223- 224
Facial recognition
, 199
-
532
Index
Folk psychology(continued
)
defined, 299
displacementof, 312, 374
Genetic
and
eliminative
and
functionalist
materialism
395
psychology
generalizations
in
299
300
" Given
399
homunculi
logical
301
304
mind
and
revision
481
383
also
Folk
of
Forel
302
Frame
395
288
Gottlob
Frege
Fritsch
Frog
16
250
Frontal
312
395
487
488n9
Prussia
162
89
165
166
201
317
340
352
349
354
and
also
pro
physical
state
Galen
"
162
352
and
Franz
263
Joseph
"
273
440
155
and
450
Ganglion
Gap
158
159
161
24
junctions
control
Gelatt
Genetics
364
367
and
, 67
of , 70
neurotransmitter
Handedness
Hanson
471
Gazzaniga
basis
release , 359
. See Dominance
- cell
Gaze
Ca + + current
Hamadryas baboon, 96
289
164
Galvanism
, " 352
cellular
Galileo
terminals ,
Habituation
80
163
combination
" Gubbins
15
Galilean
Galilei
Gall
acid
mental
( - y - aminobutyric
Phineas
, A . , 209
400
358
Functionalism
as
Grinnell
Growth
66
163
351
See
state
Greenough , W . T ., 359
354
GABA
W . , 187
251
processes
Functional
, Harold
, R . A . , 93
" Grandmother
. 9
18
359
278
, 88 , 89 , 92 . See also
callos urn
340
hormones
of
163
248
Goltz , F . L . , 161
Gonodal
28
kind
Functional
Gorski
Theodor
Functional
Gage
140
Functionalism
cesses
psychology
322
lobes
, P . 5 . , 95
Goldman
Androgen ; Estrogen
394
311
311
Great
Gustav
94
289
phenomenon
will
457
Goldman
Gordon
,
See
433
Electrical
437
Michael
. ,
Folk
Henri
Walter
Freemartin
Free
399
292
303
the
Freeman
397
444
problem
Frederick
303
301
August
Foveation
385
physics
theory
290
of
origins
383
384
displacement
as
288
networks
349
Irreducibility
theory
See
293
also
theory
310
( see
Folk
in
effects , in neuronal
461 - 464
406
processes
and
sex , 89 - 90
Global
and
brain
Glees , P . , 171
349
307
as
sex , and
. ,
Hamad
439
186
189
- evolution
284
285
, Stevan , 199
198
467
co
, N . R . , 251 , 267
synapses
Hawkins
, RobertD., 72
Head, Henry, 161
Hebb, D. 0 ., 154, 163
Index
.3
Helmholtz
254 - 255
, Hermann
von
, 22 - 24 , 32 ,
Idealism
, 246 , 249
disillusionment
savants
, 232 , 233
, R . Yorke , 174
Hillyard , S. A ., 216
Hinton , Geoffrey E., 460, 463, 468, 470,
472 , 473
Ind
uction
, in
science
, 250
relations , 344,
345
Inferior
colliculus
electrical coupling
Inferior
temporal
lobe , 172
Inferior
olive
and
functional
and memory
multicellular
in , 62
localization
, 171 - 172
, 369 , 370
studies
of , 359
Information
, 62
processing
in cell assemblies
, 381
and columnar
His
as parallel , 415
, Wilhelm
, 28
, 406 , 461
Hubel
' IInnocent
, vs . science
, 481 - 482
, Robert
A . , 471
Hydrocephalus , 106
" Hyperdopaminia ," schizophrenia as,
83
, 250
emergence
Inorganic
co - evolution
of columns
, 364 - 367
, 134 - 135
(C. S. Sherrington ), 30
Intelligence , 150, 266, 388, 395
Intelligibilia
Hyperspace , 421
Intention
449 - 450
, " 324
ions , 48
, 133- 135
Memory
Infralinguistic catastrophe, 388- 392
Inhibitory interneurons , 133- 135
Inhibitory postsynaptic potentials
Inna
, D . H . , 113 , 133
organization
533
, world
, and
behavior
534 Index
Intentionality (continued)
and
computers
Jouvet
, 342 - 346
vs . derived
mental
states , 377
term
265
Kandel
Kant
communication
, Andrew
Kimura
, D
.,
. See
70 ,
72 , 359
368
- 249
, 250
, 275
271
165
187
Accidental
kind
; Natural
Kinsboume
; Functional
kind
, Marcel
functionalism
, 356 - 358
Kitcher
Introspective
,
,
, 364
173
195
, 228
( quoted
. , 206
Edward
Knowing
, 230
modes
, 232
of
, 332
Knowledge
access
, 334
172
S . , 467
Philip
Kleitman
Klima
fluid , 48
372
Kirkpatrick
dualism
, 29
and
, 258
, 40 , 70 , 71
Intracellular
- 255
B . , 94
, Anthony
kind
and
254
, 46 , 247
Darcy
Kertesz
Kinds
conditions
of ,
R . , 66 - 67 ,
, Immanuel
Kenny
properties , 330
Intemeurons
theory
, Eric
Kelley
, 336 - 337
Interneuronal
, 271
369
. , 208
, 342 - 346
, M
Justification
to , 392
and
classical
and
logical
- 395
, 488n9
.9
empiricism
, 244
empiricism
, 258
, 267
, 270
271
reduction
, 334 - 335
encoding
fluids
of
Kohonen
Kolb
Bryan
Konishi
Reality
, Teuv
,
of ,
. V . O . Quine
" Knowledge
ogy
Involuntary
behavior , 15- 16. See also
Control , unity of
Ion pump . See Sodium -potassium pump
Ion channels , single , 48, 50- 55, 59
Ions , in cellular
sentential
and
, 394
, 265
,"
242
' o , 463
163
195
, Masakazu
Kuhn
392
264
, 94 , 445
, Thomas
S ., 6 ,
7 , 251
, 48 , 59 . See also
specificions
Ipsilateral innervation , 186, 187
IPSPs. SeeHyperpolarization
Irreducibility , 299, 312, 323, 327- 334,
349 , 378 - 379 , 487n9 . 7
Laminae
and
, Karl
geniculate
- 113
Lateral
S .,
L - dopa
120
Learning
and
. See
72 - 76 ,
,
of
173
( LGN
),
102
74 ,
75 ,
100
, 174
theories
- 201
, 216
R . I . , 359
specific
, and
nerve
energies
Parkinson
's
disease
86
LEA
118
, 459
nuclei
, 95 ,
selection
" Law
theory
171
inhibition
Latham
- 119
cognitive
passim
Lateralization
117
. , 454
paradigm
Lateral
Late
Lashley
Jellyfish, 47
Sentential
109
110
Edwin
Language
, 26 ,
Land
. See
Left
ear
ad
. See
also
Plasticity
Aplysia
Cali
/ arnica
van
tage
, 369
,"
,
21
68 - 69 ,
Index
and
co - evolution
and cognitive
, 368 - 373
Locuscoeruleusnorepinephrine system,
reclassification
in , 269
115
as term , 266
types of , 368
laws
as " virtual
governor
Leech . See Hirudo
Lesions
, of , 172
and
on behavior
interpretation
lations
localizationist
and
restoration
of , 163 - 164
model
, 171
and
causal
Lordosis
, 376 , 380
, 90
Lorento
, 331
relations
, 164 - 165
, 376 , 380
and
paradigm
specificstructures
effects
also Sentential
in , 254 , 255
reduction
defined
of , 248 - 249
, " 368
Lemniscal
535
de
Luzzatti
), A . S. F ., 167- 168,
No , R "
, Claudio
131
, 230 , 232
168
Librium
McCollum
Mach
, G "
488n10
,7
Machine
, 454
Macko
, 73 , 75 , 76
s ta tes , 304
, Kathleen
Macrocolumns
A "
220
, 136 - 137
Macromolecules
, 50 , 59
Macropotentials
, 209
Macrotheory , 6
McTaggart, J. M . E., 485n6.3
Magendie , Fran~ois, 18, 20, 21, 46
" Malsburg synapses," 477
Manipulospatial skills , 182, 190
Mapping . SeeTopographical mapping
Lobster , spiny , 78
Marie
Localization
Marmoset
388
. 7 . See also
Ten
organization
and
electrical
and
lesion
stimulation
and
, Pierre
Marr , David
, 133
, 165 - 168
, 161
, 96
, and
reduction
, 251 , 252 ,
536
Index
in functionalism
and computer
344 - 345
intentional
and
mental
and
truth
, 256
nuclei , 125
; Semantic
memory
(planet ), 262
Metabolic
processes
lum ; Pons
Midbrain
, 460
in , 149
brain
Milner
, Brenda
Mind
metaphor )
as material , 240 (see also Reduction )
Mishkin
489n9
Mitchison
.1
roles
, 182
causal
, 370
271 - 275
, 6
as ill defined
, 221
Microcolumns
memory
rehearsal
, 382
co - evolution
neurology
reversal
, 223
, Mortimer
, Graeme
, 220
, 208
Modules
and
Cognitive capacities
Mohr , J. P ., 189
consciousness
as emergent
, E . W . , 488n9 . 10
Microtheory , 6
models
, 382
between
governors
as content
, 450 - 457
of , 306- 307
relation
Mercury
Menzel
, 344
Short - term
and
as virtual
cedural
control
conditions
Medial geniculate
neurofunctionallevels
rational
378
, 344
verification
motor
and
recognition
linguistic , 337
in logical empiricism , 256- 257, 258
network theory of, 266, 337, 344
of sentences
and
tionalism )
as relational
as irreducible
, 321
Property dualism )
537
Index
Baboons; Macaque monkey ; Rhesus
monkey
Montreal Neurological Institute , 194, 201
Moral
Necker
issues
Naturalism
Natural
Neo
- Kantians
Network
Networks
, neuronal
Neurath
Neuroaxis
deafferentation
and
flocculonodular
and
lateralization
, 431 - 432
lobe , 413
, 248 - 249
coordination
and
, 152 , 354
, Otto
, 138
endogenous, 78
and
vermis
, 413
as
neurotransmitters
, 79 - 80
sex , 88 - 97
and synaptic
transmission , 69
Neuroethology , 395
Neurofunctional theory, 361, 382. Seealso
Connectionism ; Tensor network theory
laminar
structure
location
of , 166
of , 117
data
neurons
Motor
. See Efferent
neurons
as
from
term
, 153 - 154
N euromod
ula tors
Neuromuscular
, 66
junction , 28
427 , 428
Mountcastle
, Vernon
(quoted)
Neuronal
ensembles
and
coordinate
transformations
, 428 -
358 , 361
EEGs
, 209
Multiple sclerosis, 43
formation
Muscles
contraction
extraocular
of , 18 , 19 , 435 - 436
, 121 , 434
of , 477 , 478
functional
relations
between
, 427
and
and
mental
states
, 279
, 382
implants , 86- 87
, 24 - 27 , 38 - 40 . See also
neurons ; Dopaminergic
ters
en -
, 445
Neurochemicals
, 186 , 193
. See Neuron
sembles
Efferent
arrays
neurons
; Oculomotor
Neuronal
ensembles )
Afferent
neurons ;
neurons
538
Index
Neurons
(continued )
functional
between
, 27 - 30
properties
of , 41, 48- 77
development of (seeNeurogenesis)
electrophysiology of, 22, 23, 30, 48- 54,
and
habituation
and
sensi
Newton
, Sir
Nicholas
individual
, 41 , 403 , 459
Nixon
, Richard
Imaging
Nodes
as organizational
Noise
death
of , 139
of
N ottebohm
of , 42
, 59
N eurochemicals
as
term
term
, 154
Neuroscience
units
functional
neurons
, 82
( noradrenalin
) , 80 , 82
, F . , 94 , 360
( non
- rapid
eye
movement
sleep
),
Magnetic
Nuclei
Resonance
late
( NMR
) Im -
, 221
nuclei
; Medial
Numerical
Lateral
geniculate
attitudes
genicu
nuclei
, 304 , 342
theories
, 361
Observation
and
classical
and
Kant
and
logical
empiricism
empiricism
Observation
of , 267
Neurotransmitters
meaning
and
, 63 , 77 - 82 . See also
EPSPs , 64 , 80
, 254
of , 288 - 289
of , 256
self - knowledge
Occipital
lobe
Ocular
, 306 - 307
, 172
dominance
bands
, 67 - 68
270 , 274
definitions
irreducibility
level , 350
, 253
sentence
and
mechanistic
, 245 - 247
, 248
as eigenpsychische
and
, 463
. See Norepinephrine
canonical
, 470
, 154
Neuropsychology
and methodology , 147, 148- 153
as
model
, 255 , 256
, 210 , 216
Norepinephrine
aging
Neurophysiology
and biochemistry , 59
and psychology , 378- 379
Magnetic
- Sejnowski
N oradrenergic
NREM
) , 301
, 61
Hinton
waves
types
among
Ranvier
Nondetecting
Noradrenalin
variation
( president
generalizations
, 79
Imaging
, in
N100
system
. See Nuclear
Nomological
me -
, 26
NMR
dopamine
stain
Resonance
level , 359
, 302
, L . E . , 188
Nissl
of , 40
passim
. See Classical
chanics
Electrical potentials )
classification
, 70
, 286 - 293
mechanics
Nigrostriatal
functional
tion
Isaac
Newtonian
, 70 , 359
tiza
neurons
Ojemann
, George
Olfactory
system
and
ERP
technique
neural
assemblies
neural
induction
Oligodendrocytes
Ontological
, 434
, 202
, 360
of , 359 - 360
in , 38
, 41
simplification
, 280
539
Index
Ontology , 243
Opia tes, endogenous, 69
Optic chiasm, 110, 174, 176, 177
Optic tectum , 140- 144. Seealso Superior
colliculus
, 167 , 201
external vs . introspective
, 306
, 172 - 173
, M . T . , 226
, R . , 209
Ostension
, Wilder
Orangutan , 96
Organic dementia , 218, 219
Organic ions, 48
Organization , levels of, 137, 294, 356,
Orkand
Penfield
, 182
, 269
Peterhans
, E ., 471 - 472
PET Scanners
. See Positron
- Emission
To -
mography
Pain
Phantom
, 41 , 78
Parallel distributed
processing
(POP ).
See Connectionism
limbs , 16
and
multimodal
, 439
in nervous
variants
in , 462 - 463
Parallel processing
and
motor
control
neuronal
, 424 , 460
arrays , 428
postreceptor
and
brain
function
, 410
and
' s disease , 68 - 69 , 82 - 83 , 86 , 87
- ocular
.3
model , 471
classification
vestibula
in Hinton - Sejnowski
in , 455
sensory
vs . motor , 421 - 423
two - dimensional
, 441 , 443
Paralysis, 155
Parkinson
neurons
of , 79 , lIS ,
116
Phenothiazines
, 82 , 83 , 84
Phosphorus , 221
Photoreceptors, 454
Phrenology , 155, 158, 164
Physicalism, 339, 347, 352, 489n8.9. See
area 17 lesions
and
extra striate
and
motor
Pavlov
control
, Ivan
, 113 - 115
cortical
Petrovich
Peirce , Charles
also Reduction
areas , 115
, 473 - 474
and
, 253
Sanders
393
Pellionisz
, Andras
theory
Pellionisz
computer
network
brain
states
functional
behavior
, 66 - 67 , 72 , 73 , 369
in development , 40
439
Pellionisz
and
- Llinas
work theory
model
. See Tensor
net -
kinds
and
of , 151 - 152
lesions
, 413
540
Index
tensor
Proprioceptors , 46
Protoplasmic processes. SeeAxons
Prototypes, 453, 454, 455- 456
events , 66- 67
network
model
, 429 , 445
PIa to , 242
Poizner
, Howard
Polarization
Schizophrenia
Psychic pneuma, 15, 18
Psychological categories. Seealso Learning ; Memory ; Mental states; Proposi-
, 232
, 53
tional
377 , 379
attitudes
as nonnatural
Postsynaptic membrane, 64
Postsynaptic response, 52
revision
kinds
, 378
of , 383 , 384
Prefrontal
Psychosis , 68
, 66, 67
P300
and columnar
and
ocular
organization
dominance
, 133
Pure
reason , 2 , 242
429
Dualism
Procedural
waves
Purkyne , J. E., 24
Purpura , Dominick , 403 (quoted )
Pylyshyn , Zenon , 354- 355 (quoted ),
378- 379 (quoted ), 380
Pyramidal cells , 117
, 371 , 484n4 . 1
. 1 . See also
drites , processes
Axons
; Den
of
mechanics
, 253 , 286
over ; Se-
and
causation
contents
defined
of behavior
, 385
, G ., 92
, 336 - 337
Ra tionalism
value
of , 385 - 386
304 , 305
, 245
Ra tionali ty
Index
and irreducibility , 335- 336, 337
and logic, 248, 249
as term
, 266
studies
. See Cerebral
blood
flow
studies
of , 251
mind , 318
Receptive field
defined
, 120
and somatosensory
Recall
of dreams
, 206
Wada
reduc -
tion
and
co - evolution
defined
, 278
domain
- relative
, 364 - 367
, 357
retentive
, as
term
, 278
, 429
, mechanistic
17
Refractory period , 58
Rehearsal buffer , 149
account
of , 15 - 16 ,
541
542
Index
Richards
Whitman
Right
ear
Right
hemisphere
Robotics
488n10
Irvin
Roger
471
( crab
( quoted
See
Semantic
memory , 372
Semicircular
canals , 434
) ,
196
197
Lateralization
Sensa tions
. 2
Roffwarg
235
~ REA
Rock
advantage
472
. ,
- like
207
208
robot
) ,
420
433
472
Sense , certainties
and
coordinate
and
phase
442
transformation
sandwich
Sense - data
446
hypothesis
Russell
David
Bertrand
Institute
Salmon
. ,
251
82
41
and
61
83
85
218
and
,
in
Science
Scientific
256
258
and
Reality
( Wilfrid
of
and
folk
and
functionalist
theory
in
psychology
in
and
neurobiological
and
unified
295
301
realism
Scientific
reasoning
. ,
operations
Searle
See
squirt
Sejnowski
472
See
295
311
mental
Sentential
at -
Manifesto
, 386 - 387
487n8
.9
252
475
and
attentional
476
and
brain
and
connectionism
in visual
. 9
Serotonin
Terrence
. ,
Hinton
impairment
function
463
468
Sejnowski
, 410
, 461
and somatosensory
Califarnica
487n8
also
, 473
349
13
Selective
298
249
Aplysia
John
370
472
models
hypothesis
hare
296
312
Searchlight
connectionist
titudes
theory
Scientific
for , 222
295
theory
Scoville
psychology
generalizations
Sea
psychology
absence
Sea
254
281
Perception
Sellars
of , 70 - 71
and synchronicity , 62
Sensory aphasia. SeeWernicke ' s aphasia
Sensory cortex, 117, 439
Sensory neurons . SeeAfferent neurons
Sensory phase space, 422, 427- 428. See
253
pseudoscience
revolution
as device
basis
419
61
system
, cellular
control
220
Science
logical
, 46 - 47
. See Observation
, 242
Sensitization
Sensorimotor
336
294
conduction
cell
252
brain
,
Schizophrenia
Schwann
Sensibilia
458
230
Wesley
Saltatory
and
sentences
sentence
Senses , of animals
,
Salk
space
443
Rumelhart
as
of , 250
473
470
model
system
system , 129
, 113
, 70 - 71 , 79 , 80 , 115
, and
neurochemicals
88 - 87 I 483n2 .4
410
Self
as
and
ill
defined
tion , 207
182
of
173
178
- knowledge
Sellars
Sellars
Semantic
181
neuropharmacology
unity
Self
Wilfrid
Feyerabend
issues
182
305
383
90 , 96 - 97
223
310
301
thesis
69
Short - term
302
301
302
memory
, 202
543
Index
SimplicityJ as theoreticalcriterionJ263
Sleep, 204- 209, 206. SeealsoNREM
sleep; REMsleep
Smart, J. J. C., 271
Smith, Aaron, 188- 189
Snakevenom, and synapse, 67
Sodiumamytal. SeeWadatest
Sodiumions (Na+), 48, 51- 55, 56, 58
Sodium-potassiumpump, 48, 49
Sokolofftechnique, 484n5.2
Soma, 38, 49, 54
Somatosensory
cortex, 127, 129, 133
Somatosensory
evokedpotentials(SEP),
210, 212
Somatosensory
system, 46, 108, 110,
127- 130
Somatotopicmaps, 127- 130
Soul, 180, 182, 243, 316, 320
Space
, representationof, 120, 232
Spatiotemporalintegration, 200
Specialization
, in nervoustissue, 171,
173
Specificity, response
, 43, 45
in animals, 46- 47
in cortex, 113, 159, 160
of individual neurons, 459
and Kant, 46
Speech
and Broca's area, 159, 189, 259- 260
and Brodmannarea6, 201- 202
comprehensionof, 157
lesionsaffecting, 157
motor control for, 193
Spencer
, Herbert, 162, 484n4.4 (quoted)
Sperry, R. W., 140, 174- 176, 189, 325,
487n9.1
Spinalcord, 90, 105, 131
Spinothalamicsystem, 108, 110, 129
Split-brain studies, 174- 193
and unity of consciousness
, 321
Split personality, 181- 182
Sprague,J. M., 115
Springer, SallyP., 197
Spurzheim, JohannKaspar, 158
Squid, 32, 38, 62
Squire, L. R., 360, 371
Staining, and nervecells, 24. Seealso
Carminestain; Golgi staining method;
Nissl stain
Statespace. SeePhasespacerepresenta
tion
State vectors
, 460
Statocysts, 47
Stereoptic vision , 121, 125
Stich, Stephen, 382, 383, 386, 457
Stimulus
of , 20 - 21
Stumpf , Walter, 92
Subcognitive
processes , 149
Subjective experience
and
antireductionism
and
brain
states
, 317
, 326 , 327
, 129
Substance
dualism
and
Descartes
and
Plato
, 242
Substantia
Sulci
, 243
extraocular
and
tensorial
laminar
muscles
, 131
transformations
architecture
, 439
of , 444
visual
attentive
processes
, 478
synapses
, 40
communication
types of , 63
structure
, 30
544 Index
Synaptic connections, 52
Synaptic delay, 64
Synaptic junctions , 30, 31
Synaptic potentials , 51- 54
Synaptic transmission , 80
Synaptic weighting , 464, 471
Synaptogenesis, 54
Synchronicity , 62
Synthetic spatial processing, 199- 200
Systematicity, and theory , 467, 489n10
Theories
reduction
Thermodynamics
and
co - evolution
, 363
and cooperative
Tachistoscopic studies
and asymmetries
and
in normal brains ,
, 251
Techniques
for charting pathways , 107- 108
for fixation
, 27
reduction
, 356 - 358
.6
constraints
imagin
- ~,
- 217- 222
staining , 24- 27
for three dimensional
Tracers , radioactive
, 203 - 210
371 , 449
scanning , 84
transformation
Terminal
, Anne , 475
common
idiom
, 271 - 274
, 265
definitional
, 254
Testosterone
Thalamus
losum
Treisman
Truth
and
cessmg
Tensor
, 218
, and
bulbs
brain , 89 - 91
, 100 , 102
and columnar
sentential
organization , 131
electrophysiology in , 476
and language functions , 202
redprocal connections involving , 121- 123
reticular complex of, 476
somatotropic
mapping
in , 127
Truth
, 251
conditions
, 344
cortex
cells , 113
Index
to
Type
transformation
type
identity
Unconditioned
Unified
to
and
co
and
displacement
of
375
goal
of
and
psychology
and
reductionism
Valium
Heerden
Van
Wagenen
Vecchi
Vector
10
277
249
. ,
Vitalism
407
174
417
488n10
. 1
model
of , 113
, 461
106
107
of
485n5
Marie
Arouet
, 18
. 6
meaning
\ ois
behavior , 16- 18
See
tion
theory
, Fran
Voluntary
449
transforma
45
256
257
VOR
. See Vestibulo
- ocular
reflex
267
413
Vertebrate
brains
Vesalius
Vesicles
Wada
37
Andreas
nuclei
434
Vestibulocerebellum
Vestibulo
ocular
Watts , J. W . , 163
439
reflex
( VOR
108
109
Weiskrantz
Wernicke
111
gaze
control
sensorimotor
of
network
Vienna
Circle
Virchow
Virtual
theory
433
Wernicke
438
, "
365
366
368
373
374
Visual
agnosia
agnosia
Visual
phase
Visual
processing
recognition
AI
224
456
225
See
also
as
420
hierarchical
421
White
matter
Wiesel
, T . N . , 113 , 133
Willshaw
Color
agnosia
space
115
Visual
Facial
' s area
41
governors
, 226
441
253
Rudolf
, Lawrence
, Carl
433
transformation
tensor
39
62
Vestibular
and
in , 474 - 478
, 15 , 18 , 22 , 29
Voltaire
418
transformation
te
Vermis
"
mechanisms
176
467
vector
Verification
and
, 104
279
"
learning
Ventricles
and
reflexes
315
mathematics
Coordina
10
activity
to
of , 149
hierarchical
10
William
. ,
matrix
cognitive
265
and
tests
Visual
maps
,
cerebellar
, 473 - 474
contours , 471
Visual system
286
82
and
Vector
brain
261
Van
control
motor
of subjective
attentional
Reality
Uranus
and
71
293
Unknowable
457
passim
of
220
285
shape
. ,
mind
impossibility
"
evolution
general
as
374
443
Leslie
theory
blocks
354
stimulus
ngerleider
441
545
112
, 102 , 103
, David
, 461
Wilson
, Donald
, 176
Wilson
, Mark , 470
, Fo , 372
546
Index
WordandObject(W. V. O. Quine), 2
Word blindness, 164
Word recognition, 454, 459- 460
" World 3." SeeIntelligibilia, world of
Xenon 133, 218
Yin/Yang, and lateralization , 183
Young , J, Z " 349 (quoted ), 481
Zaidel
, Eran
, 184
finches
, 94
lateralization
186 , 185
, Steven
W " 471
studies, 184-